# 3D AI Studio Documentation (full) > Complete documentation for 3D AI Studio, a browser-based platform for generating 3D models, textures, and images using AI. No software install needed. 3D AI Studio lets users create 3D models from text prompts or images, generate and edit images with 15+ AI models, create textures and materials, rig characters, and export to all major 3D formats (GLB, FBX, OBJ, STL, USDZ, BLEND, 3MF). Everything runs in the browser at 3daistudio.com. AI assistants can also use 3D AI Studio directly through its MCP connector at https://mcp.3daistudio.com/mcp. --- # Getting Started Source: https://docs.3daistudio.com/getting-started Welcome! This quickstart gets you from zero to a finished 3D model in about five minutes. Everything runs in your browser - no downloads, no installs, no credit card required. [Video: 3D AI Studio Full Guide](https://www.youtube.com/watch?v=MUp-dMf8mLU) ## 1. Create your account Go to the [AI 3D model generator at 3daistudio.com](https://www.3daistudio.com) and click **Get Started**. Sign up with email or Google - takes about 10 seconds. You get free credits immediately, so you can jump straight into generating. ![3D AI Studio signup modal with email and social login options](https://docs.3daistudio.com/assets/images/docs/signup-modal.jpg) *The signup modal with email and social login options for creating your 3D AI Studio account.* ## 2. Your Dashboard After signing in you land on the Dashboard. This is home base - every model you generate shows up here as a card you can preview, download, or process further. ![3D AI Studio dashboard showing tools and recent generations](https://docs.3daistudio.com/assets/images/docs/dashboard.jpg) *The 3D AI Studio dashboard showing your tools, recent generations, and model cards.* Look at the top-left corner for your **credit balance**. Credits are what you spend each time you generate a model. The free credits you got at signup are enough to try several generations. ![Credit balance display in 3D AI Studio dashboard](https://docs.3daistudio.com/assets/images/docs/credits-section.jpg) *The credit balance display in the top-left corner of the dashboard.* ## 3. Your first Text to 3D model Let's generate something. Click **Text to 3D** in the sidebar (or open the [Text to 3D model generator](https://www.3daistudio.com/TextTo3D) directly). 1. **Write a prompt** - Describe what you want. Keep it specific: *"a medieval treasure chest with gold trim"* works better than *"a chest"*. Mention materials, style, and details. 2. **Pick a model** - You'll see these options: - **Prism 3.1** - Best quality, 35 credits, 3-5 minutes. The default and the right choice for most projects - **Tripo P1 / P2 - Game Ready** - Clean low-poly topology in about a minute, 60 / 70 credits - **Hunyuan 3.1 Pro** - Sharp, even meshes at a flat 40 credits 3. **Hit Generate** and wait for your model to appear. Start with Prism 3.1 at Standard texture quality to iterate on prompt wording, then switch to Ultra or Max texture quality for your final version. You may also spot a Texture Model setting on Prism 3.1 - Latest is an experimental new texturing pass at the same price, and results can vary while it's being evaluated. **Quick prompt tips:** - Be specific about materials: *"brushed steel"*, *"weathered wood"*, *"matte ceramic"* - Mention style: *"low-poly"*, *"realistic"*, *"stylized cartoon"* - See the full [Prompt Tips](https://docs.3daistudio.com/3d-generation/prompt-tips) guide for more ## 4. Your first Image to 3D model Got a photo, sketch, or AI-generated image? You can turn it directly into a 3D model. Click **Image to 3D** in the sidebar (or open the [Image to 3D model generator](https://www.3daistudio.com/ImageTo3D) directly). 1. **Upload your image** - Drag it in or click to browse. For best results, use a clear shot of a single object on a simple background. 2. **Choose your AI model** - Same options as Text to 3D (Prism 3.1, Tripo P1/P2, Hunyuan 3.1 Pro and more). ![Model selection grid](https://docs.3daistudio.com/assets/images/docs/model-selection.jpg) *The model selection grid showing available AI models for Image to 3D generation.* 3. **Click Generate** and watch the progress bar. ![Generation in progress](https://docs.3daistudio.com/assets/images/docs/generating-progress.jpg) *A generation in progress, showing the progress bar and estimated time.* **Multi-image mode:** Upload photos from multiple angles (front, side, back) and the AI combines them for a more accurate reconstruction. Great for real objects you've photographed. ![Multi-image upload settings](https://docs.3daistudio.com/assets/images/docs/multi-image-settings.jpg) *The multi-image upload mode for combining photos from multiple angles into one 3D model.* ## 5. Preview and export Once generation finishes, your model appears in the built-in 3D viewer. Rotate by dragging, zoom with the scroll wheel. ![3D viewer with completed model](https://docs.3daistudio.com/assets/images/docs/viewer-result.jpg) *The interactive 3D viewer showing a completed model you can rotate, zoom, and download.* Happy with the result? Download it in any format: | Format | Best for | |--------|----------| | **GLB** | Web, AR, general use (default) | | **FBX** | Unity, Unreal Engine, Maya | | **OBJ** | Blender, ZBrush, universal | | **STL** | 3D printing | | **USDZ** | Apple AR (iPhone, iPad, Vision Pro) | | **BLEND** | Native Blender projects | | **3MF** | Color 3D printing | Click the **Download** button on your model card and pick the format you need. That's it - your 3D model is ready to use. ## 6. Next steps You've got the basics down. Here's where to go from here: - **[Image Studio](https://docs.3daistudio.com/image-studio/overview)** - Generate and edit reference images before converting to 3D. This two-step workflow (perfect image first, then convert) gives you the most control over your final model. - **[Prompt Tips](https://docs.3daistudio.com/3d-generation/prompt-tips)** - Write better prompts, get better models. - **[Texture Generator](https://docs.3daistudio.com/textures/texture-generator)** - Add or swap PBR textures on any model. - **[Workflow Examples](https://docs.3daistudio.com/3d-generation/recommended-workflow)** - Full start-to-finish workflows for games, printing, e-commerce. - **[Export & Formats](https://docs.3daistudio.com/export-formats)** - Detailed breakdown of every format and when to use it. - **[MCP (AI Assistants)](https://docs.3daistudio.com/mcp/overview)** - Generate models straight from Claude, ChatGPT, Cursor, Claude Code or Codex. Questions? Join the [Discord community](https://discord.gg/uA3u6J7EDC) or email [support@3daistudio.com](mailto:support@3daistudio.com). ## Frequently Asked Questions ### Is 3D AI Studio free to use? Yes, you can start for free. 3D AI Studio gives you free credits when you sign up, enough to generate several 3D models without entering a credit card. After your free credits are used, you can purchase more or subscribe to a plan for ongoing use. ### How many free credits do you get with 3D AI Studio? New accounts receive free credits on signup that let you try multiple generations right away. The exact amount may vary, but it is typically enough for 3-5 model generations depending on which AI model you choose. 3D AI Studio shows your credit balance in the top-left corner of the dashboard. ### Do I need 3D modeling experience to use 3D AI Studio? No experience needed at all. 3D AI Studio runs entirely in your browser - just type a description or upload an image and the AI generates a complete 3D model for you. There is no software to install and no technical knowledge required. The whole process takes about 2-5 minutes from signup to downloading your first model. ### What are the system requirements for 3D AI Studio? 3D AI Studio runs in your web browser, so you do not need a powerful computer. Any modern browser (Chrome, Firefox, Safari, Edge) with WebGL support works. A stable internet connection is recommended since the AI processing happens on the server side, not on your device. ### Which AI model should I pick for my first generation? Start with Prism 3.1 (35 credits, best quality, 3-5 minutes) - it's the default and the safest choice for a first generation. Iterate on your prompt at Standard texture quality, then switch to Ultra or Max for the final production model. If you need clean low-poly topology for a game, pick Tripo P1 or P2 instead. ### What is the difference between Text to 3D and Image to 3D? Text to 3D generates a model from a written description like "a medieval treasure chest with gold trim." Image to 3D converts a photo, sketch, or illustration into a 3D model. In 3D AI Studio, Image to 3D generally produces more accurate results when you have a specific look in mind, while Text to 3D is great for quick concept exploration. ### Can I use 3D AI Studio models commercially? Yes. Models you generate with 3D AI Studio are yours to use in commercial projects, including games, apps, product visualization, and 3D printing. Check the terms of service on 3daistudio.com for the full details on licensing and usage rights. ### How is 3D AI Studio different from other AI 3D generators? 3D AI Studio offers multiple AI models to choose from (Prism, Tripo P Series, Hunyuan, Meshy, Rodin, Trellis), built-in tools for texturing, rigging, remeshing, and format conversion, plus direct integrations with Blender, Unity, Unreal Engine, and more. It is an all-in-one platform rather than just a single generation tool. You can go from prompt to game-ready asset without leaving the site. ### What file formats can I download my models in? 3D AI Studio supports 9 export formats: GLB, FBX, OBJ, STL, USDZ, BLEND, 3MF, Collada (.dae), and PLY. GLB is the default and works almost everywhere. Use FBX for game engines, STL for 3D printing, and USDZ for Apple AR. See the Export Formats page for a detailed comparison. --- # Recommended Workflow Source: https://docs.3daistudio.com/3d-generation/recommended-workflow The best 3D models don't come from typing a prompt into Text to 3D. They come from creating a good reference image first, then converting that image to 3D. This two-step approach gives you way more control over the final result. This page shows you exactly how to do it, with real examples. ## Why Image First? When you use Text to 3D directly, the AI has to imagine what your object looks like based on words alone. Sometimes it nails it, sometimes it doesn't. And if the result is close but not quite right, you have to rewrite the prompt and try again. With the image-first workflow, you can: - **See the result before committing to 3D** - if the image doesn't look right, edit it - **Control exactly what the model looks like** - colors, style, angle, details - **Use photos of real objects** - not just text descriptions - **Edit yourself into the scene** - turn a selfie into a figurine, LEGO character, etc. The 3D conversion step is the same either way (Image to 3D with Prism 3.1). The difference is how good your input is. ## Workflow 1: Generate an Image, Then Convert to 3D This is the most common workflow. You create a clean reference image in Image Studio, then send it to Image to 3D. ### Steps 1. Open [Image Studio](https://www.3daistudio.com/ImageGeneration) and go to **Generate** 2. Pick a model (ImageGen 4 for general use, Flux 2 for photorealism) 3. Write a prompt describing what you want, with "clean white background" and "single object" 4. Generate until you get an image you like 5. Open [Image to 3D](https://www.3daistudio.com/ImageTo3D) 6. Upload the generated image 7. Pick **Prism 3.1** and generate ### Real example: Owl character Here's what this looks like in practice. We generated an owl character image, then converted it to 3D: ![Owl character reference image generated in 3D AI Studio Image Studio](https://docs.3daistudio.com/assets/images/docs/workflow-owl-input.jpg) ![3D model of owl character generated with Image to 3D](https://docs.3daistudio.com/assets/images/docs/workflow-owl-3d.jpg) Generated with "stylized 3D cartoon owl, soft lighting, clean white background", then converted to 3D with Prism 3.1. ### More examples ![Warrior character concept art as input for Image to 3D](https://docs.3daistudio.com/assets/images/docs/workflow-warrior-input.jpg) ![3D model of warrior character generated from concept art](https://docs.3daistudio.com/assets/images/docs/workflow-warrior-3d.jpg) ![Sneaker product photo used as input for AI 3D generation](https://docs.3daistudio.com/assets/images/docs/workflow-sneaker-input.jpg) ![3D model of sneaker generated from product photo](https://docs.3daistudio.com/assets/images/docs/workflow-sneaker-3d.jpg) The pattern is the same every time: generate a clean image with one subject, simple background, good lighting, then convert to 3D. ### Tips for generating good reference images - Add **"clean white background"** or **"studio lighting"** to your prompt - Include **"single object, centered"** to keep it simple - Use **"3D render style"** or **"stylized 3D"** for images that convert better than flat illustrations - Generate a few variations and pick the best one before converting to 3D - If the image is almost right, edit it in Image Studio before converting (see Workflow 2) ## Workflow 2: Edit a Photo, Then Convert to 3D Already have a photo? You can edit it into something that converts to 3D really well. This is especially useful for personalized figurines, product models, or turning real objects into 3D. ### Steps 1. Open [Image Studio](https://www.3daistudio.com/ImageGeneration) and go to **Edit** 2. Upload your photo 3. Pick **GPT-Image-2 Edit** or **Gemini Edit** 4. Write an edit prompt (e.g. "turn this person into a LEGO minifigure, plastic material, clean background") 5. Download the edited image 6. Open the [image to 3D model generator](https://www.3daistudio.com/ImageTo3D) and upload it 7. Generate with **Prism 3.1** ### Real example: Photo to LEGO figurine to 3D print This is the full pipeline. We took a regular photo, edited it into a LEGO-style figurine, converted to 3D, and 3D printed the result: ![Original photo before AI LEGO figurine conversion](https://docs.3daistudio.com/assets/images/docs/workflow-photo-input.jpg) ![Photo edited into LEGO minifigure style using AI Image Studio](https://docs.3daistudio.com/assets/images/docs/workflow-lego-edited.jpg) ![3D model of LEGO figurine generated with Prism 3.1](https://docs.3daistudio.com/assets/images/docs/workflow-lego-3d.jpg) ![3D printed LEGO figurine made from AI-generated model](https://docs.3daistudio.com/assets/images/docs/workflow-lego-printed.jpg) Photo, AI edit, 3D model, 3D print. The whole thing takes about 5 minutes (not counting print time). ### Real example: Generated LEGO orc You can also combine Workflow 1 and 2. Here we generated a LEGO orc from scratch in Image Studio, then converted to 3D: ![AI-generated LEGO orc](https://docs.3daistudio.com/assets/images/docs/workflow-orc-generated.jpg) ![LEGO orc 3D model](https://docs.3daistudio.com/assets/images/docs/workflow-orc-3d.jpg) The prompt was something like "LEGO minifigure of an orc warrior, green skin, armor, plastic material, studio lighting, white background" using ImageGen 4. ### Good edit prompts for 3D conversion | What you want | Edit prompt | | --- | --- | | LEGO figurine of yourself | "turn this person into a LEGO minifigure, plastic material, blocky proportions, full body, clean background" | | Pixar-style character | "turn into a Pixar-style 3D character, stylized, big eyes, clean background" | | Funko Pop figurine | "turn me into a Funko Pop figurine, full body, clean background" | | Clean product shot | "remove the background, keep the product exactly as is, white background" | | Stylized 3D version | "turn into a stylized 3D render, clean lighting, white background" | ## Workflow 3: Sketch to 3D Have a hand-drawn sketch or concept drawing? You can turn it into a full 3D model by rendering it first in Image Studio. ### Steps 1. Open [Image Studio](https://www.3daistudio.com/ImageGeneration) and go to **Edit** 2. Upload your sketch (photo of a drawing, digital sketch, or rough concept) 3. Use **GPT-Image-2 Edit** or **Gemini Edit** with a prompt like "render this sketch as a detailed 3D model, stylized, clean background" 4. The AI turns your rough sketch into a polished, render-ready image 5. Upload that rendered image to [Image to 3D AI](https://www.3daistudio.com/ImageTo3D) 6. Generate with **Prism 3.1** ### Real example: Orc character from sketch A rough pencil sketch becomes a fully rendered character, then a 3D model: ![Hand-drawn orc character sketch as input for AI rendering](https://docs.3daistudio.com/assets/images/docs/workflow-sketch-input.jpg) ![Orc sketch rendered into detailed 3D character by AI](https://docs.3daistudio.com/assets/images/docs/workflow-sketch-rendered.jpg) ![3D model generated from rendered orc sketch using Image to 3D](https://docs.3daistudio.com/assets/images/docs/workflow-sketch-3d.jpg) Sketch, AI render, 3D model. The rough pencil drawing was uploaded to Image Studio Edit with the prompt "render this sketch as a 3D model, stylize". The AI produced the detailed character render. That render was then uploaded to Image to 3D with Prism 3.1 to get the final 3D model. ### Tips for sketch to 3D - **Clean line art works best** - clear outlines with minimal construction marks - **Add basic shading** to your sketch if possible - it helps the AI understand depth and volume - **Three-quarter angles** convert better than straight front views - **Digital sketches** tend to work better than photos of paper drawings (less noise) - If your sketch is on paper, photograph it with even lighting and crop tightly around the subject ## Workflow 4: Direct Photo to 3D (Quick and Simple) If you just want a quick 3D model from an existing photo without editing, you can skip Image Studio and go straight to Image to 3D: 1. Open [Image to 3D](https://www.3daistudio.com/ImageTo3D) 2. Upload your photo 3. Pick Prism 3.1 4. Generate This works well for: - Product photos with clean backgrounds - Concept art and illustrations - Toys and objects photographed on a table - Character art and drawings It works less well for: - Photos with busy backgrounds (edit first to remove the background) - Multiple objects in one image (crop to just one) - Very dark or blurry photos (edit first to improve quality) ## When to Use Text to 3D Instead Text to 3D is still useful when: - You want a quick concept and don't care about exact appearance - You're exploring ideas and want fast iterations - The object is simple enough to describe in words (a sword, a potion bottle, a chair) - You don't have any reference image at all But for anything where the look matters, the image-first workflow will save you time and give you better results. ## Quick Reference | Situation | Workflow | Time | | --- | --- | --- | | I have an idea but no image | Generate image in Image Studio, then Image to 3D | ~5 min | | I have a photo of myself | Edit photo in Image Studio, then Image to 3D | ~5 min | | I have a sketch or drawing | Render sketch in Image Studio Edit, then Image to 3D | ~5 min | | I have a product photo | Upload directly to Image to 3D (or remove background first) | ~3 min | | I have concept art | Upload directly to Image to 3D | ~3 min | | I want a quick rough model | Text to 3D directly | ~3 min | | I want to 3D print a figurine | Edit photo to figurine style, Image to 3D, export STL | ~10 min | [Video: Image to 3D Tutorial](https://www.youtube.com/watch?v=OVKl7idiyIU) *Watch the full Image to 3D workflow in action.* [Open Image Studio](https://www.3daistudio.com/ImageGeneration) - [Open Image to 3D](https://www.3daistudio.com/ImageTo3D) - [Open Text to 3D](https://www.3daistudio.com/TextTo3D) ## Frequently Asked Questions ### Why is image-first better than text to 3D directly? When you type a prompt into Text to 3D, the AI has to guess what your object looks like. With the image-first workflow in 3D AI Studio, you see the exact reference image before spending credits on 3D generation. You can edit colors, fix proportions, remove backgrounds, and iterate cheaply on the 2D image until it's perfect. This leads to far more predictable 3D results. ### How many credits does the full image-first workflow cost? A typical run in 3D AI Studio costs around 40-50 credits total. Image generation is about 5-10 credits depending on the model, and Image to 3D with Prism 3.1 costs 35 credits. That's slightly more than Text to 3D alone (25-60 credits), but you get much more control over the outcome. Image editing adds another 5-10 credits if needed. ### What is the best workflow to generate a 3D model with AI? The best approach in 3D AI Studio is the image-first workflow. Instead of going straight to Text to 3D, generate a reference image first in Image Studio, edit it until it looks exactly right, then convert it to a 3D model using Image to 3D. This gives you much more control over the final result because you can perfect the 2D image cheaply before committing credits to 3D generation. Image generation costs 2-8 credits while 3D generation costs 25-60, so iterating in 2D saves both time and credits. ### Can I turn a hand-drawn sketch into a 3D model? Yes, and it works surprisingly well. Upload your sketch to Image Studio Edit in 3D AI Studio and use a prompt like "render this sketch as a detailed 3D model, clean background." The AI turns your rough drawing into a polished render, which you then convert to 3D with Image to 3D. Clean line art with some basic shading converts best. ### What's the best image generation model for 3D conversion? ImageGen 4 in 3D AI Studio works great for stylized and cartoon objects. Flux 2 is better for photorealistic subjects. Both produce images that convert well to 3D. The key is adding "clean white background" and "single object, centered" to your prompt regardless of which model you pick. ### How do I turn a selfie into a 3D figurine? Upload your photo to Image Studio Edit in 3D AI Studio and use a prompt like "turn this person into a LEGO minifigure, plastic material, full body, clean background." Once the edited image looks good, send it to Image to 3D with Prism 3.1. Export as STL for 3D printing. The entire process takes about 5 minutes. ### Does this workflow work for game assets? Absolutely. Many game developers use 3D AI Studio's image-first workflow to rapidly prototype assets. Generate a concept in Image Studio with style keywords like "low-poly" or "game asset," then convert to 3D with Prism 3.1 or Tripo P1 Game Ready. The output can go straight into Unity or Unreal after a quick retopology pass if needed. ### What image backgrounds work best for 3D conversion? Solid white or transparent backgrounds give the best results in 3D AI Studio. Studio lighting setups also work well. The goal is to isolate your subject clearly so the AI doesn't create unwanted geometry from background elements. If your source image has a busy background, use Image Studio's background removal tool before converting. ### Can I combine workflows for complex objects? Yes, many users mix approaches in 3D AI Studio. For example, you might generate a character's body with the image-first workflow, then generate accessories separately with Text to 3D, and combine everything in Blender. You can also generate multiple angle references in Image Studio and use a multi-view model for more accurate reconstruction. ### How long does the entire recommended workflow take? From start to finished 3D model, the image-first workflow in 3D AI Studio typically takes about 5 minutes of active work. Image generation takes seconds, editing (if needed) takes a minute or two, and Prism 3.1 generates the 3D model in 3-5 minutes. Direct photo to 3D (without editing) is even faster at about 3 minutes total. ### What if my generated image isn't perfect? That's the beauty of the image-first approach. In 3D AI Studio, you can edit the generated image before converting to 3D. Fix colors, remove unwanted details, change the background, or adjust proportions using Image Studio Edit. Each edit costs only 5-10 credits, which is much cheaper than regenerating the entire 3D model if something is off. --- # Image to 3D Source: https://docs.3daistudio.com/3d-generation/image-to-3d Image to 3D turns a photo, sketch or illustration into a textured 3D model. Upload the image, pick an AI model and download the result as GLB, FBX, OBJ, STL or USDZ, usually within 1-10 minutes depending on the model. | | | | --- | --- | | **Input** | One PNG, JPG or WebP image, or several views of the same object | | **Cost** | 15-75 credits depending on the model, 35 for the default Prism 3.1 | | **Time** | 60 seconds to 12 minutes, 3-5 minutes for Prism 3.1 | | **Output** | Textured mesh with optional PBR maps | | **Downloads** | GLB, FBX, OBJ, STL, USDZ | [Video: Image to 3D Tutorial](https://www.youtube.com/watch?v=OVKl7idiyIU) ## How do I turn a picture into a 3D model? Open [Image to 3D](https://www.3daistudio.com/ImageTo3D), upload the picture, choose a model such as Prism 3.1 and click **Generate**. The AI reconstructs the full shape, including the sides it cannot see, and textures it with the colors from your image. 1. Open Image to 3D from the sidebar or use the [photo to 3D model converter](https://www.3daistudio.com/ImageTo3D) on 3daistudio.com 2. Add your image by dragging it onto the upload area, clicking to browse, or pasting it with Ctrl+V (Cmd+V on Mac) 3. Check the preview thumbnail to confirm the image looks right 4. Choose an AI model (see the comparison below) 5. Optionally adjust texture quality, mesh quality, PBR materials or face limit 6. Click **Generate**, then inspect and download the model in the 3D viewer ![Model picker showing available Image to 3D models](https://docs.3daistudio.com/assets/images/docs/model-selection.jpg) *The model picker in Image to 3D, showing all available AI models with their credit costs and quality tiers.* ## Which AI model should I choose? Start with **Prism 3.1**: it is the default and the best balance of quality, speed and cost. Switch models only for a specific need such as game-ready topology, speed or a lower price. | If you want... | Use | Why | | --- | --- | --- | | The best all-round result | **Prism 3.1** | Default, state-of-the-art quality for 35 credits | | The fastest result | **Tripo P1 Game Ready** | Finishes in 60-90 seconds | | Game assets | **Tripo P2 Game Ready** | Optimized topology and optional quad meshes | | Maximum detail | **Hunyuan 3.1 Pro** | State-of-the-art detail and textures | | The lowest price | **Trellis 2** | Starts at 15 credits | | The newest texturing (experimental) | **Prism 3.1** with Texture Model set to **Latest** | New texturing pass with a Fast tier and Neutral Lighting | ## How do the single-image models compare? All single-image models build the complete 3D shape from one photo. They differ in price, speed and detail. | Model | Credits | Speed | Quality | Time | | --- | --- | --- | --- | --- | | **Prism 3.1** (default) | 35 | Standard | State of the art | 3-5 min | | Hunyuan 3.1 Pro | 40 | Standard | State of the art | 3-4 min | | Tripo P2 Game Ready | 70 | Fast | State of the art | 1-3 min | | Tripo P1 Game Ready | 60 | Fast | State of the art | 60-90 sec | | Meshy 7.1 | 40 | Standard | State of the art | 5-10 min | | Trellis 2 | 15-41 | Standard | High | 4-8 min | | Rodin 2.5 | 30 | Fast-Standard | State of the art | 6-7 min | | Forge Gen-2 | 35 | Standard | State of the art | 5-7 min | | Hunyuan 3.0 | 40 | Standard | State of the art | 3-4 min | | Prism 3.0 (legacy) | 30 | Standard | High | 2-4 min | | Seed3D 1.0 | 30 | Standard | High | 8-12 min | Seed3D 1.0 is in beta and available on paid plans only. Prism 3.1's Latest texture model costs the same as Standard. Its results can vary while it is being evaluated, so Standard remains the recommended choice for production work. ![Generation in progress with status indicator](https://docs.3daistudio.com/assets/images/docs/generating-progress.jpg) *A model generation in progress, showing the real-time status indicator and estimated time remaining.* ## When should I use a multi-view model? Use a multi-view model when you have photos of the same object from several angles and the hidden sides must be accurate, for example product models or characters whose back matters. Single-image models guess the unseen sides; multi-view models reconstruct them from real views. There are two input types: - **Fixed views:** the model expects specific angles (front, left, back, right) and you assign each image to its slot. - **Flexible multi-image:** upload several images from any angles and the model works out how they relate. Multi-image models are available on paid plans only. | Model | Credits | Input | Time | | --- | --- | --- | --- | | **Prism 3.1 Multi-view** | 40 | 4 fixed views (front, left, back, right) | 3-5 min | | Hunyuan 3.1 Pro Multi-view | 45 | 1-8 assigned views | 3-4 min | | Tripo P2 Multi-view | 75 | 4 fixed views | 1-3 min | | Tripo P1 Multi-view | 65 | 4 fixed views | 60-90 sec | | Meshy 7.1 Multi-image | 45 | 1-4 images | 5-10 min | | Rodin 2.5 Multi-image | 35 | 2-5 images | 6-7 min | | Forge Gen-2 Multi-image | 40 | 2-5 images | 5-7 min | | Hunyuan 3.0 Multi-view | 45 | 4 views | 3-4 min | | Prism 3.0 Multi-view (legacy) | 35 | 4 fixed views | 2-4 min | Keep the same lighting, distance and background in all photos. Inconsistent shots confuse the reconstruction. ![Multi-image upload interface with view assignment](https://docs.3daistudio.com/assets/images/docs/multi-image-settings.jpg) *The multi-image upload interface, where you assign photos to specific viewpoints for more accurate 3D reconstruction.* ## What makes a good input image? A sharp, well-lit image of a single object on a plain background gives the best 3D model. Busy backgrounds, harsh shadows and heavy compression turn into unwanted geometry or noise. 1. **Plain background:** solid white, gray or transparent works best. Remove a cluttered background first with the one-click tool in [Image Studio](https://docs.3daistudio.com/image-studio/utilities). 2. **Even lighting:** avoid harsh shadows and overexposed areas. 3. **Sharp, high-resolution image:** blur and compression artifacts show up as bumps in the mesh. PNG gives the cleanest result. 4. **Subject fills the frame:** centered and not cropped. 5. **Slight 3/4 angle:** showing the front and one side gives more depth cues than a flat front view. Phone photos work well when the object is in focus and well lit. If a photo has strong shadows or a color cast, set the Texture Model to **Latest** (Prism 3.1 or Tripo P2) and turn on **Neutral Lighting**. ## What do the generation settings do? The settings control texture resolution, polygon density and the extra maps you get. The defaults suit most images, so change them only when you have a specific target. - **Texture quality:** resolution of the texture maps. Higher is sharper and takes slightly longer. - **Mesh quality:** polygon density. Higher gives smoother surfaces and finer detail. - **Mesh detail (Meshy 7.1):** Meshy's Ultra geometry pass. Standard, Ultra 2K (+10 credits) or Ultra 4K (+15 credits, single image only). - **PBR materials:** adds normal, roughness and metallic maps to the base color, for realistic rendering in game engines. - **Face limit:** the maximum polygon count, useful for specific platforms. - **Seed:** the same image, seed and settings reproduce the same result. - **Auto-size:** scales the model to realistic real-world dimensions. - **Texture Model (Prism 3.1, Tripo P2):** Standard (default) or Latest (experimental), which adds the Fast texture quality and Neutral Lighting at the same price. - **Neutral Lighting:** removes shadows, highlights and color casts baked into the photo before texturing. Available with the Latest texture model. - **Quad Mesh Output (Tripo P2):** four-sided polygons for rigging and subdivision, delivered as FBX alongside a GLB preview. No extra cost, capped at 25,000 polygons. ## Which formats can I download? Download the finished model as **GLB, FBX, OBJ, STL or USDZ** directly from the 3D viewer. PBR maps are included as separate image files with OBJ and FBX. | Format | Use for | | --- | --- | | **GLB** | Blender, Unity, Unreal, web viewers, general use | | **FBX** | Game engines and animation pipelines | | **OBJ** | Static meshes, maximum compatibility | | **STL** | 3D printing (geometry only) | | **USDZ** | AR Quick Look on Apple devices | More formats and comparisons are in [Export Formats](https://docs.3daistudio.com/export-formats). ![3D viewer showing a generated model with download options](https://docs.3daistudio.com/assets/images/docs/viewer-result.jpg) *The built-in 3D viewer displaying a completed model, with format options for downloading.* ## What's next [Turn your picture into a 3D model](https://www.3daistudio.com/ImageTo3D) now, with free credits on signup. To start from a description instead of a picture, see the [Text to 3D guide](https://docs.3daistudio.com/3d-generation/text-to-3d). For the cleanest results, follow the [recommended workflow](https://docs.3daistudio.com/3d-generation/recommended-workflow): perfect the reference image first, then convert it. ## Frequently Asked Questions ### Is there a free option for image to 3D? New accounts get free credits on signup, enough to test several generations. After that, models cost 15 credits (Trellis 2) up to 75 credits (Tripo P2 Multi-view). There is no free unlimited tier. ### Can I 3D print a model made from a photo? Yes. Open the model in [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing), which makes it watertight, sets the height in millimeters and downloads an STL, or a multi-color 3MF for an AMS or MMU. Solid, chunky objects print better than thin, delicate ones. The full walkthrough is in our article on [how to 3D print a picture](https://www.3daistudio.com/blog/how-to-3d-print-a-picture-from-2d-to-3d-in-easy-steps). ### Does image to 3D keep the colors from my photo? Yes. Every model textures the mesh with the colors and details from your image. Enable PBR materials to also get normal, roughness and metallic maps. Texture sharpness depends on the input resolution and the texture quality setting. --- # Text to 3D Source: https://docs.3daistudio.com/3d-generation/text-to-3d Text to 3D generates a textured 3D model from a written description. Type a short prompt, pick an AI model and download the result as GLB, FBX, OBJ, STL or USDZ, usually within 1-10 minutes and without any modeling skills. | | | | --- | --- | | **Input** | A text prompt, ideally 5-20 words about one object | | **Cost** | 30-70 credits depending on the model, 35 for the default Prism 3.1 | | **Time** | 60 seconds to 10 minutes, 3-5 minutes for Prism 3.1 | | **Output** | Textured mesh with optional PBR maps | | **Downloads** | GLB/GLTF, FBX, OBJ, STL, USDZ | [Video: Text to 3D Tutorial](https://www.youtube.com/watch?v=Y05LgsWPjoY) ## How do I generate a 3D model from text? Open [Text to 3D](https://www.3daistudio.com/TextTo3D), describe the object in one short sentence, choose a model such as Prism 3.1 and click **Generate**. When it finishes, inspect the model in the 3D viewer and download it. 1. **Write a prompt** describing one object, its materials and style 2. **Pick a model** based on quality, speed or budget 3. **Adjust settings** such as texture quality, face count or PBR (optional) 4. **Click Generate**. Processing time depends on the model 5. **Preview the result**: rotate, zoom and inspect it in the 3D viewer 6. **Download** in the format you need ![3D viewer showing a generated model result](https://docs.3daistudio.com/assets/images/docs/viewer-result.jpg) *A completed 3D model displayed in the interactive viewer, ready for export.* ## How do I write a good text to 3D prompt? Describe one object in 5-20 words and include its material, style and view. Short, focused prompts produce cleaner models than long paragraphs or scenes with several objects. - **Be specific:** name the object, its materials and key features - **Mention the view:** "full body", "side view" or "isometric" helps the framing - **Add a style:** "realistic", "low-poly", "cartoon" or "PBR" guide the look - **One object per prompt:** "a sword and a shield" tends to merge into one messy mesh. Generate them separately and combine them in your 3D software | Use case | Prompt that works well | | --- | --- | | Architecture | "medieval castle with stone walls and a tall tower" | | Character | "cute chibi anime cat girl, big eyes, full body" | | Weapon or prop | "futuristic sci-fi pistol, metallic, detailed" | | Vehicle | "red sports car, glossy paint, side view" | More examples are in [Prompt Tips](https://docs.3daistudio.com/3d-generation/prompt-tips). ## Which model should I choose? Start with **Prism 3.1**, the default with the best balance of quality, speed and cost. Switch only for a specific need. | If you want... | Use | Why | | --- | --- | --- | | The best overall quality | **Prism 3.1** or **Hunyuan 3.1 Pro** | State-of-the-art geometry and textures | | The fastest result | **Tripo P1** or **Tripo P2** | About one minute | | Game-ready assets | **Tripo P2 Game Ready** | Clean topology, optional quad meshes | | The lowest price | **Rodin 2.5** | From 30 credits | | Maximum detail | **Prism 3.1** with Ultra mesh, **Hunyuan 3.1 Pro** or **Forge Gen-2** | Best for complex objects | | The newest texturing (experimental) | **Prism 3.1** with Texture Model set to **Latest** | New texturing pass with a Fast tier | ![Model selection grid in 3D AI Studio](https://docs.3daistudio.com/assets/images/docs/model-selection.jpg) *The model selection grid in 3D AI Studio, showing available Text to 3D models with credit costs and quality ratings.* ## How do the text to 3D models compare? | Model | Credits | Speed | Quality | Time | | --- | --- | --- | --- | --- | | **Prism 3.1** (default) | 35 | Standard | State of the art | 3-5 min | | **Hunyuan 3.1 Pro** | 40 | Standard | State of the art | 3-4 min | | **Tripo P2 Game Ready** | 70 | Fast | State of the art | 1-3 min | | **Tripo P1 Game Ready** | 60 | Fast | State of the art | 60-90 sec | | **Meshy 7.1** | 40 | Standard | State of the art | 5-10 min | | **Rodin 2.5** | 30 | Fast-Standard | State of the art | 6-7 min | | **Forge Gen-2** | 35 | Standard | State of the art | 5-7 min | | **Hunyuan 3.0** | 40 | Standard | State of the art | 3-4 min | | **Prism 3.0** (legacy) | 30 | Standard | High | 2-4 min | ![3D model generation in progress](https://docs.3daistudio.com/assets/images/docs/generating-progress.jpg) *A 3D model generation in progress with the real-time status indicator.* ## Which settings can I change? Every model family has its own options for texture quality, polygon count and output. The defaults work for most prompts. ### Prism - **Texture quality:** Standard, Ultra (+5 credits) or Max (+10 credits) - **Mesh quality:** Standard or Ultra (+10 credits) - **PBR:** physically based rendering textures - **Face limit:** up to 1,500,000 faces (2,000,000 with Ultra mesh quality) - **Texture Model (Prism 3.1):** Standard (default) or Latest (experimental) at the same price. Latest adds a Fast texture tier (same price as Standard, quicker, less fine detail) ### Hunyuan - **Generation type:** Normal, Geometry Only, Sketch or LowPoly - **PBR:** full PBR material maps - **Max face count:** from 40,000 up to 1,500,000 faces ### Meshy 7.1 and Meshy 6 - **Mesh detail (Meshy 7.1):** Standard, Ultra 2K (+10 credits) or Ultra 4K (+15 credits) - **Mesh quality:** Standard or Smart Low-Poly (AI) for clean game-ready topology up to 15,000 polygons (20 credits with textures) - **Texture quality:** Standard (2K), Ultra (4K, +10 credits) or Max (8K, +15 credits) - **Topology:** triangle or quad mesh - **Remesh option:** cleaner topology for animation - **Character pose:** A-pose or T-pose - Meshy 7 stays available as a legacy version; Meshy 5 has been retired ### Tripo P1 and P2 Game Ready - **Texture on or off:** P1 costs 60 credits with texture and 40 without; P2 costs 70 with texture and 50 without - **Quad Mesh Output (P2):** four-sided polygons for rigging and subdivision, delivered as FBX alongside a GLB preview. No extra cost, up to 25,000 faces - **Texture Model (P2):** Standard (default) or Latest (experimental), which unlocks the Fast texture quality ### Rodin 2.5 - **Quality tiers:** from extreme-low to extreme-high - **HighPack add-on:** 4K textures for +10-15 credits ## Should I use Text to 3D or Image to 3D? Use Text to 3D for quick models straight from an idea. Use Image to 3D when the look matters, because you can see and refine the reference image before spending credits on 3D. > [!NOTE] > For maximum control, generate an image first in [Image Studio](https://docs.3daistudio.com/image-studio/overview) with the same prompt, edit it until it looks right, then convert it with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d). Image generation is cheaper, so you can iterate freely before committing credits to 3D. The full process is in the [recommended workflow](https://docs.3daistudio.com/3d-generation/recommended-workflow). ## Which formats can I download? Download the model as **GLB/GLTF** for web and AR, **FBX** for Unity and Unreal, **OBJ** for maximum compatibility, **STL** for 3D printing or **USDZ** for Apple AR. GLB is the safest all-rounder. The [Export Formats guide](https://docs.3daistudio.com/export-formats) compares them in detail. [Try the Text to 3D generator](https://www.3daistudio.com/TextTo3D), with free credits on signup. ## Frequently Asked Questions ### How much does text to 3D cost per model? 30 to 70 credits. Rodin 2.5 and Prism 3.0 start at 30 credits, the default Prism 3.1 costs 35, and Tripo P2 Game Ready costs 70. Options such as Ultra texture quality (+5) or Ultra mesh quality (+10) add a few credits. New accounts get free credits to test several models. ### Can AI generate game-ready 3D models from text? Yes. Tripo P1 and P2 Game Ready produce clean, game-optimized topology in about one to three minutes, and P2 can output quad meshes. For other models, run the [Remesh tool](https://docs.3daistudio.com/processing/remesh) or [Smart Low Poly](https://docs.3daistudio.com/processing/smart-low-poly) afterward to prepare the mesh for Unity or Unreal Engine. ### How do I get PBR textures on text to 3D models? Turn on the PBR option before generating. Prism and Hunyuan models output base color, normal, roughness and metallic maps. With FBX or OBJ downloads, the PBR maps come as separate texture files. ### What is the maximum polygon count? Up to 1.5 million faces with Prism 3.1 (2 million with Ultra mesh quality) and Hunyuan 3.1 Pro. Most other models produce around 40,000-100,000 faces. Set a face limit before generating to match your platform: lower for mobile and web, higher for film and renders. ### What is the Latest texture model and should I use it? Latest is a newer, experimental texturing pass on Prism 3.1 and Tripo P2 at the same price as Standard. It adds a Fast texture tier and, for image inputs, Neutral Lighting. Results can vary while it is being evaluated, so use it to experiment and stay on Standard for production work. The geometry is identical either way. --- # SVG to 3D: Turn Vector Files into 3D Models Source: https://docs.3daistudio.com/3d-generation/svg-to-3d The SVG to 3D tool takes any SVG file and extrudes it into a 3D model - right in your browser. Everything runs client-side, so your files never leave your computer. It's completely free, no credits or signup required. It works great for logos, icons, typography, and brand assets. You get a live 3D preview with full material and lighting controls, then export as GLB (3D model) or PNG (transparent screenshot). ![SVG to 3D upload area](https://docs.3daistudio.com/assets/images/docs/svg-empty-state.jpg) *The SVG to 3D upload area, ready for you to drag and drop an SVG file.* ## How to Use It **1. Upload your SVG** Drag and drop an SVG file onto the upload area, or click to browse. Max file size is 5MB. **2. Preview in 3D** Your SVG appears as an extruded 3D model with a live preview you can rotate, zoom, and pan. The tool preserves your original SVG colors by default. ![Loaded SVG with 3D preview](https://docs.3daistudio.com/assets/images/docs/svg-loaded-1.jpg) *An SVG file loaded and extruded into a 3D model with a live interactive preview.* **3. Adjust settings** Use the settings panel to tweak thickness, materials, lighting, and more. Changes show up instantly in the preview. **4. Export** Download your result as a GLB file for use in 3D apps, games, or AR - or grab a transparent PNG for presentations and marketing materials. ![Another SVG to 3D example](https://docs.3daistudio.com/assets/images/docs/svg-loaded-2.jpg) *Another SVG to 3D example showing a different design with customized material and lighting.* ## Settings The settings panel has five tabs. Here's what each one does: ![SVG to 3D settings panel](https://docs.3daistudio.com/assets/images/docs/svg-settings.jpg) *The settings panel with tabs for geometry, material, environment, background, and view controls.* **Geometry** Controls the shape of the extrusion. - **Thickness** (1-250, default 4) - How deep the extrusion goes. Higher values give a chunkier, more solid look. - **Bevel size** (0-5, default 0.3) - Rounds off the edges. A small bevel adds realism; zero gives you sharp edges. - **Bevel segments** (1-64, default 4) - How smooth the bevel curve is. More segments = smoother rounding. **Material** Pick a look or build your own. - **Presets** - Choose from Metallic, Clay/Matte, Plastic, Rough Glass, Glass, or Custom - **Roughness** - How shiny or matte the surface is - **Metalness** - Metal vs non-metal appearance - **Clearcoat** - Adds a glossy clear layer on top (like car paint) - **Transmission** - Makes the material see-through (for glass effects) - **Override SVG colors** - Replace the original SVG colors with a single material color **Environment** Sets the lighting around your model. - **HDRI presets** - City, Dawn, Forest, Lobby, or Studio lighting - **Custom color** - Use a solid color background for lighting instead of an HDRI - **Intensity** - How bright the environment lighting is - **Bloom** - Adds a soft glow effect around bright areas **Background** Pick a background color for the viewport using the color picker. **View** - **Auto-rotate** - Slowly spins the model so you can see all sides - **Rotation speed** - How fast it spins ## Material Presets Not sure which material to use? Here's a quick guide: | Preset | Best for | |--------|----------| | **Metallic** | Logos, badges, premium brand assets | | **Clay/Matte** | Clean mockups, 3D printing previews | | **Plastic** | Icons, playful designs, product viz | | **Rough Glass** | Frosted or etched effects | | **Glass** | Transparent overlays, decorative pieces | | **Custom** | Full control over every slider | ## Export Options Both export options run entirely in your browser - nothing gets uploaded. - **GLB** - A standard 3D file format. Works in Blender, Unity, Unreal, web viewers, and AR apps. This is your go-to if you need the actual 3D model. - **PNG** - A transparent-background screenshot of the current view. Great for slide decks, social media, and marketing materials where you just need a good-looking render. ## Tips for Best Results **Which SVGs work best** - Simple, clean paths with solid fills - logos, icons, and text work great - SVGs with well-defined shapes and clear outlines extrude cleanly - Bold, thick designs look better in 3D than very thin, delicate linework **Getting cleaner output** - Start with a small thickness value and increase gradually - Add a slight bevel (0.2-0.5) to make edges catch light naturally - If your SVG has many tiny details, simplify it in a vector editor first - Try different HDRI presets - lighting makes a big difference in how materials read **Common issues** - Very complex SVGs with hundreds of paths may render slowly - simplify them if you can - Overlapping paths can create odd geometry - flatten or merge paths in your vector editor before uploading - If colors look wrong, check whether "Override SVG colors" is turned on in the Material tab ## Use Cases - **Logos for 3D websites** - Export a GLB of your logo and embed it in a Three.js or model-viewer scene - **Icons for games** - Extrude UI icons or item symbols for use in Unity or Unreal - **Brand assets for AR** - Turn brand marks into AR-ready GLB files for product packaging or experiences - **Social media content** - Export a styled PNG render of your logo for posts and thumbnails - **3D printing** - Extrude a design, export the GLB, and import it into your slicer ## What's Next Ready to try it? Head to the [SVG to 3D Converter](https://www.3daistudio.com/svgTo3D) and drop in an SVG. If you want to generate 3D models from images instead, check out the [Image to 3D guide](https://docs.3daistudio.com/3d-generation/image-to-3d). For text-based generation, see [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d). ## Frequently Asked Questions ### Is the SVG to 3D converter really free? Yes, completely free with no signup, no credits, and no limits. The SVG to 3D tool in 3D AI Studio runs entirely in your browser, so nothing gets uploaded to a server. You can convert as many SVG files as you want and export GLB or PNG files without any account. ### Can I convert a logo to a 3D model with this tool? Absolutely. Logos are one of the best use cases for the SVG to 3D tool in 3D AI Studio. Just export your logo as an SVG from Illustrator, Figma, or any vector editor, then upload it. The tool extrudes it into 3D with customizable depth, materials, and lighting. Export as GLB for web 3D viewers or AR experiences. ### What's the difference between GLB and PNG export? GLB gives you an actual 3D model file you can import into Blender, Unity, Unreal, or any web-based 3D viewer. PNG gives you a transparent screenshot of the current view, which is great for presentations, social media, or mockups. Both exports in 3D AI Studio's SVG to 3D tool are free and run locally in your browser. ### Why does my SVG look wrong after extrusion? Complex SVGs with overlapping paths or very thin strokes can produce odd geometry. Before uploading to 3D AI Studio, try flattening or merging overlapping shapes in your vector editor. Also check if "Override SVG colors" is accidentally enabled in the Material tab, which replaces your original colors with a single uniform color. ### Can I use the exported GLB for 3D printing? Yes, but you'll need to convert the GLB to STL first. Export the GLB from 3D AI Studio's SVG to 3D tool, then import it into Blender or an online converter to save as STL. For best printing results, use a bold font or thick-stroked design, increase the extrusion depth to at least 5, and add a slight bevel so edges aren't razor-sharp. ### What SVG file size and complexity limits should I know about? The max file size is 5MB. In practice, most logos and icons in 3D AI Studio's SVG to 3D tool are well under this. However, very complex SVGs with hundreds of tiny paths (like detailed illustrations) may render slowly or produce messy geometry. For best results, keep your SVG simple with clean, well-defined shapes and merge unnecessary paths before uploading. --- # 3D Text & Logo Maker Source: https://docs.3daistudio.com/3d-generation/3d-text The [3D text generator](https://www.3daistudio.com/font3d) turns any text into an extruded 3D model directly in your browser. Pick a font, adjust the depth and material, and export as GLB or PNG. Everything runs locally - nothing gets uploaded to a server. Free, no credits, no signup. ![3D Text default view](https://docs.3daistudio.com/assets/images/docs/3d-text-default.jpg) *The 3D Text tool with default settings, showing extruded text in the live preview.* ## How It Works 1. **Type your text** - enter any text in the text field. Supports multiple lines (use Shift+Enter for new lines). 2. **Pick a font** - choose from 16 built-in fonts or upload your own TTF/WOFF2 file. 3. **Adjust the look** - change depth, bevel, materials, lighting, and colors in real time. 4. **Export** - download as GLB (3D model) or PNG (transparent render). The 3D preview updates instantly as you change settings. Rotate, zoom, and pan the model with your mouse. ![3D Text example with custom text](https://docs.3daistudio.com/assets/images/docs/3d-text-example.jpg) *Custom text rendered in 3D with adjusted font, depth, and material settings.* ## Settings The settings panel has 7 tabs: ### Text Type your text and choose alignment (left, center, right). Multi-line text is supported. ### Typography ![Typography settings](https://docs.3daistudio.com/assets/images/docs/3d-text-settings.jpg) *The typography settings panel with controls for font, size, depth, letter spacing, and bevel.* | Setting | Range | What it does | | --- | --- | --- | | Font | 16 built-in + custom upload | Choose the typeface | | Size | 1-50 | How large the text appears | | Depth | 0.5-100 | How far the text extrudes into 3D | | Line Height | 0.5-3 | Spacing between lines | | Letter Spacing | 0-10 | Gap between characters | | Curve Segments | 1-64 | Smoothness of curved letters | **Bevel** (optional) adds rounded edges to the text. When enabled: - Bevel Thickness (0-2) - Bevel Size (0-2) - Bevel Segments (1-16) - higher = smoother bevel ### Available Fonts Roboto, Open Sans, Bahiana, Caveat, Doto, Doto Bold, Doto Rounded, Kablammo, Lobster, Nova Mono, Rationale, Silkscreen, Silkscreen Bold, Space Grotesk, Space Grotesk Bold, Train One Want a different font? Click "Upload your own" and select any TTF or WOFF2 file from your computer. ### Geometry - **Spread** (0-100) - expands or contracts the shapes - **Invert Depth** - flips the extrusion direction - **Invert Mask** - cuts the text shape out of a background rectangle ### Material ![Different material applied to 3D text](https://docs.3daistudio.com/assets/images/docs/3d-text-material.jpg) *3D text with a different material preset applied, showing how materials change the overall look.* Five presets to choose from: | Preset | Look | Best for | | --- | --- | --- | | Metallic | Shiny metal surface | Logos, premium branding | | Clay / Matte | Soft, non-reflective | Prototypes, clean renders | | Plastic | Slight sheen, subtle reflection | Product mockups | | Rough Glass | Frosted translucent | Artistic, decorative | | Glass | Clear transparent | Special effects, overlays | Or go custom with individual sliders for roughness, metalness, clearcoat, and transmission. You can also override the text color with any custom color. ### Environment Choose how the scene is lit: - **HDRI mode** - pick from 5 environment presets: City, Dawn, Forest, Lobby, Studio - **Color mode** - set a single light color - Adjust intensity (0-3) - Optional bloom effect for a soft glow ### Background Pick any background color. Default is dark (`#121212`). Reset to default with one click. ### View - Auto-rotate toggle with adjustable speed (0.1-5) ## Export Two export options, both free: | Format | What you get | Best for | | --- | --- | --- | | GLB | 3D model file with materials | Web viewers, AR, game engines, 3D printing | | PNG | Transparent screenshot of the current view | Social media, presentations, marketing | Export buttons appear in the top-right corner of the preview panel. ## Tips - **For 3D printing** - use a bold font with thick strokes (Silkscreen Bold, Space Grotesk Bold). Increase depth to at least 5. Export GLB and convert to STL using the [Quad Remesh Tool](https://www.3daistudio.com/Tools/Remesh). - **For web/AR** - keep curve segments low (4-8) for smaller file sizes. GLB works directly in most web 3D viewers. - **For transparent PNGs** - set the background to a dark color, adjust lighting for the look you want, then export as PNG. The background will be transparent. - **Multi-line text** - use Shift+Enter to add line breaks. Adjust line height to control spacing. - **Custom branding** - upload your brand font (TTF/WOFF2), set your brand colors, and use the metallic or plastic preset for a professional look. ## Use Cases - **Social media headers** - create 3D text renders for YouTube thumbnails, Instagram posts, Twitter banners - **Logo mockups** - extrude your logo text in 3D with different materials and lighting - **Game UI** - generate 3D text for game titles, menus, and HUD elements - **Presentations** - add 3D text renders to slides for visual impact - **3D printing** - print custom name plates, signs, and decorative text [Open the 3D Text and Logo Maker](https://www.3daistudio.com/font3d) ## Frequently Asked Questions ### Is the 3D text generator completely free? Yes, 100% free with no signup, no credits, and no usage limits. The 3D Text tool in 3D AI Studio runs entirely in your browser, so your text never leaves your computer. You can create and export as many GLB or PNG files as you want without paying anything. ### Can I use my own custom font for 3D text? Yes. Click "Upload your own" in the typography settings and select any TTF or WOFF2 font file from your computer. 3D AI Studio's 3D Text tool renders it instantly in 3D with all the same customization options as the built-in fonts. This is great for brand-specific typography. ### How do I export 3D text for 3D printing? Export as GLB from 3D AI Studio's 3D Text tool, then convert to STL using a tool like Blender or the Quad Remesh tool on the platform. For best printing results, use a bold font with thick strokes (like Silkscreen Bold or Space Grotesk Bold), set the depth to at least 5, and increase curve segments for smoother letters. ### Can I create multi-line 3D text? Yes. Press Shift+Enter to add line breaks in the text input field. 3D AI Studio's text tool renders each line in 3D, and you can adjust the line height setting to control spacing between lines. Text alignment (left, center, right) applies to all lines. ### What's the best material preset for logo mockups? Metallic works best for premium, corporate logo presentations. It catches light beautifully and gives a polished, professional feel. For a softer look, try Clay/Matte in 3D AI Studio's 3D Text tool. Plastic adds a subtle sheen that works well for playful or consumer brands. You can also go fully custom with individual sliders for roughness, metalness, and clearcoat. ### Can I use the GLB file in web-based 3D viewers? Yes. GLB is the standard format for web 3D and works directly with Three.js, model-viewer, Babylon.js, and most web-based viewers. Export your 3D text from 3D AI Studio, upload the GLB to your server, and embed it in any webpage. Keep curve segments low (4-8) for smaller file sizes that load faster on the web. --- # 3D Relief Generator Source: https://docs.3daistudio.com/3d-generation/relief-generator The Relief Generator turns any photo into a 3D relief sculpture - the kind of raised surface you see on coins, decorative panels, and carved plaques. Upload an image, and AI converts it into a depth map that becomes printable 3D geometry. The output works for 3D printing, CNC milling, and laser engraving. ![Relief Generator idle state with floating example reliefs](https://docs.3daistudio.com/assets/images/docs/relief-empty-state.jpg) *The Relief Generator landing view with floating example relief sculptures.* ## How It Works The tool uses a two-stage process: 1. **Depth map generation (server-side)** - AI analyzes your photo and produces a grayscale height map where bright areas become raised surfaces and dark areas stay flat. This step costs credits. 2. **Mesh building (client-side)** - Your browser instantly converts that depth map into a 3D mesh you can rotate, adjust, and export. No extra credits needed for tweaking settings. This means you can regenerate the depth map once and then experiment with dozens of different 3D settings for free. ## Choosing a Depth Style Before generating, pick a depth style that matches your image: - **Detailed** - Best for objects, logos, architecture, and anything with sharp edges or fine surface detail. Preserves crisp lines and small features. - **Organic** - Best for portraits, animals, and smooth natural shapes. Produces gentler depth transitions that avoid harsh artifacts on skin and curved surfaces. If you're unsure, try Detailed first. Switch to Organic if the result looks too noisy or harsh on smooth areas. ## AI Models | Model | Credits | Max Resolution | Best For | |-------|---------|---------------|----------| | **Nano Banana 2** | 4 | 2K | General use, fast results | | **Gemini Edit** | 4 | 2K | Fine detail, complex scenes | | **Nano Banana Pro** | 6-12 | 4K | High-res prints, large reliefs | Nano Banana Pro supports 4K resolution but doubles in cost (12 credits) at that setting. For most projects, 1K or 2K at 4 credits is plenty. ![Upload panel showing AI model picker and depth style options](https://docs.3daistudio.com/assets/images/docs/relief-settings.jpg) *The upload panel with AI model picker, depth style options, and resolution settings.* ## 3D Settings Explained After generation, you get a live 3D preview with real-time controls. Settings are grouped by function: ### Mesh Mode - **Solid** - Creates a watertight mesh with a back plate. Use this for 3D printing where your slicer needs a closed volume. - **Planar** - Surface only, no back. Use for CNC carving paths or laser engraving files where you just need the relief surface. ### Depth Controls - **Depth Intensity** (0-0.3) - How tall the relief is. Start around 0.1 for subtle reliefs, go higher for dramatic depth. - **Depth Cutoff** (0-95%, planar mode only) - Removes background geometry below a threshold. Useful for isolating a subject from its background. - **Base Thickness** (0.01-0.5, solid mode only) - How thick the back plate is. Thicker = stronger print but more material. - **Invert for Mold** - Flips the depth so raised areas become recessed. Use this when making casting molds or stamp negatives. ### Smoothing and Cleanup - **Smoothing** (0-20px) - Blurs the depth map before mesh generation. Removes noise but also softens detail. - **Noise Reduction** (0-50) - Targets random speckling without blurring edges as much. Recommended 15-30 for portraits. - **Surface Smoothing** (1-10 passes) - Laplacian smoothing applied to the final mesh. Higher passes = smoother polygons. - **Enhance Details** (paid only) - Amplifies fine surface detail. Good for architectural elements and text. ### Quality and Display - **Mesh Quality** (128-1024) - Polygon resolution of the output. Higher values capture more detail but produce larger files. - **Show Texture** - Toggles the original image mapped onto the mesh surface. - **Material Presets** - 12 options including Matte White PLA, Metallic Bronze, Glossy Resin, Ceramic, and more. These are for preview only and don't affect the exported file. ![Generated relief sculpture shown in the 3D viewer](https://docs.3daistudio.com/assets/images/docs/relief-generated-1.jpg) *A generated relief sculpture displayed in the 3D viewer with adjustable depth and material controls.* ## Solid vs Planar Mode **Use Solid when:** - 3D printing (FDM or resin) - slicers need watertight geometry - Creating standalone decorative pieces - Making lithophanes (thin relief panels that show images when backlit) **Use Planar when:** - CNC milling - you only need the carving surface, not the stock material - Laser engraving - flat relief paths without unnecessary underside geometry - Combining the relief with other 3D models in external software ![Another relief example showing depth and detail](https://docs.3daistudio.com/assets/images/docs/relief-generated-2.jpg) *Another relief example highlighting the depth detail and surface texture of the generated output.* ## Export Options Download your relief in two formats: - **STL** - Geometry only. The standard for 3D printing and CNC. Works with every slicer and CAM program. - **OBJ** - Includes geometry plus optional texture as a ZIP. Use when you want the original image mapped onto the relief surface. Export requires a paid plan. Free users can generate and preview reliefs with full 3D controls, but downloading is locked until you upgrade. ![Export panel showing STL and OBJ download buttons](https://docs.3daistudio.com/assets/images/docs/relief-export-panel.jpg) *The export panel with STL and OBJ download options for 3D printing and CNC use.* ## Tips for Best Results 1. **High-contrast images work best** - Clear separation between subject and background produces cleaner depth maps. 2. **Portraits** - Use Organic depth style, noise reduction at 20-30, and smoothing at 5-10px. This prevents skin from looking bumpy. 3. **Logos and text** - Use Detailed depth style with minimal smoothing. Bump up Enhance Details if available. 4. **Large prints** - Use Nano Banana Pro at 4K and mesh quality 512+. The extra resolution prevents visible stair-stepping on big pieces. 5. **Lithophanes** - Solid mode, thin base (0.05-0.1), low depth intensity (0.05-0.08). Print in white PLA or resin for best light transmission. 6. **Avoid busy backgrounds** - Crop your image tight around the subject, or the relief will include unwanted background geometry. ## Common Use Cases - **3D printed wall art** - Turn family photos into framed relief panels - **Lithophanes** - Thin relief prints that reveal an image when you hold them up to light - **CNC carvings** - Mill decorative panels, signs, or architectural details from wood or aluminum - **Custom gifts** - Pet portraits, wedding photos, or logos as physical keepsakes - **Mold making** - Use Invert for Mold to create negatives for casting in resin or chocolate - **Laser engraving** - Planar mode exports work directly as depth paths for laser engravers ## What's Next Ready to make your first relief? Head to the [3D Relief Generator](https://www.3daistudio.com/Tools/3DReliefGenerator) and upload a photo. For full 3D models from images (not just reliefs), check out [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d). ## Frequently Asked Questions ### How much does it cost to generate a 3D relief? Depth map generation in 3D AI Studio starts at just 4 credits using Nano Banana 2 or Gemini Edit. Nano Banana Pro costs 6 credits at standard resolution and 12 credits at 4K. Once you have the depth map, all 3D mesh adjustments (height, smoothing, resolution, materials) are free since they run in your browser. ### Can I make a lithophane with this tool? Yes, the Relief Generator in 3D AI Studio works great for lithophanes. Use Solid mode with a thin base (0.05-0.1), low depth intensity (0.05-0.08), and export as STL. Print in white PLA or translucent resin for the best effect. When you hold the thin print up to a light source, the image appears through varying thickness. ### What image types work best for 3D relief generation? High-contrast photos with clear separation between subject and background produce the cleanest reliefs. Portraits, logos, architectural shots, and product photos all work well in 3D AI Studio's Relief Generator. Avoid very dark images, low-contrast scenes, or photos where the subject blends into the background. ### Should I use Solid or Planar mode for CNC carving? Use Planar mode for CNC milling. It generates only the relief surface without a back plate, which matches how CNC carving works - you're cutting into existing stock material. 3D AI Studio exports the planar mesh as STL, which you can import directly into your CAM software (Fusion 360, VCarve, etc.) as a carving toolpath surface. ### Can I create a mold from a relief for resin casting? Yes. Enable "Invert for Mold" in 3D AI Studio's Relief Generator before exporting. This flips the depth so raised areas become recessed cavities, giving you a negative you can print and use as a casting mold. Works well for resin, chocolate, soap, and other pourable materials. ### What resolution should I use for large 3D prints? For large prints (over 15cm), use Nano Banana Pro at 4K resolution with mesh quality set to 512 or higher in 3D AI Studio. This prevents visible stair-stepping on the relief surface. The extra resolution costs 12 credits instead of 4, but it's worth it for pieces where you'll see fine surface detail up close. --- # Prompt Tips Source: https://docs.3daistudio.com/3d-generation/prompt-tips This page covers how to write prompts that actually work for three things: generating 3D models from text, generating images, and editing existing images. No theory, just practical tips and real examples. ## Text to 3D Prompts When you write a prompt for [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d), the AI needs to understand what object to build, what it looks like, and what style you want. Keep it between 5 and 20 words for most cases. Be specific but don't overload it. ### The formula **Object + key features + material/style + constraints** ### Prompts that work | Prompt | What you get | Good for | | --- | --- | --- | | "medieval castle with stone walls and a tall tower" | Detailed architectural model | Environment assets | | "cute chibi anime cat girl, big eyes, full body, pastel colors" | Stylized character | Game characters, figurines | | "futuristic sci-fi pistol, metallic, detailed, hard surface" | Clean weapon model | Game props | | "red sports car, glossy paint, side view, realistic" | Vehicle with good proportions | Product visualization | | "wooden stool, three legs, worn oak, subtle wood grain, one solid piece" | Clean furniture model | Interior design | | "fantasy potion bottle, chunky glass, wide base, glowing liquid inside" | Stylized prop | RPG game assets | | "sci-fi helmet, smooth curved shell, dark matte polymer, symmetrical" | Hard surface model | Character equipment | | "cute chubby robot mascot, round body, simple shapes, friendly face" | Cartoon character | Branding, mascots | ### What makes prompts better - **Name the material** - "wooden", "metallic", "glass", "plastic", "stone" gives the AI clear texture direction - **Mention the style** - "realistic", "stylized", "low-poly", "cartoon", "photorealistic" sets the aesthetic - **Say "one solid piece"** - prevents the AI from generating floating parts or disconnected geometry - **Add "clean geometry"** - helps with printing and game engine compatibility - **Specify view** - "front view", "side view", "three-quarter view" for more predictable results - **Use real names** - "wingback chair" works better than "chair with wings on the sides" ### What to avoid - **Too vague** - "a cool thing" or "something nice" gives the AI nothing to work with - **Too long** - paragraphs of text can confuse the model. 5-20 words is the sweet spot - **Contradictions** - "realistic cartoon" or "simple with extreme detail" sends mixed signals - **Multiple objects** - "a sword and a shield and a helmet" usually produces a messy result. Generate one object at a time - **Scene descriptions** - "a knight standing in a forest next to a dragon" is a scene, not a single 3D model. Focus on one object ### Model-specific tips | Model | Works best with | Tip | | --- | --- | --- | | Prism 3.1 | Detailed objects, realistic style | Add "PBR texture" for better materials | | Tripo P1 | Quick concepts, simple low-poly objects | Keep prompts shorter (5-10 words) | | Hunyuan 3.1 Pro | Organic shapes, characters | Try the "Geometry" mode for untextured sculpts | | Meshy 6 | Characters, creatures | Use "full body" and specify a pose | | Forge Gen-2 | Detailed props, hard surface | Mention specific materials and surface finish | | Tripo P1 | Game-ready assets | Add "game ready" or "game asset" to the prompt | ## Image Generation Prompts When generating images in [Image Studio](https://www.3daistudio.com/ImageGeneration), different AI models need different prompting styles. But all of them respond well to clear, descriptive language. ### The formula **Subject + action/pose + style + context/lighting** ### Prompts that work | Prompt | Model | What you get | | --- | --- | --- | | "stylized 3D cartoon owl, soft lighting, clean white background" | ImageGen 4 | Clean reference for Image to 3D | | "professional product photo of a leather watch, studio lighting, white background" | Flux 2 | Product photography | | "anime warrior girl with blue hair and katana, full body, cel shaded" | ImageGen 4 | Character concept art | | "LEGO minifigure of an orc warrior, plastic material, studio lighting" | ImageGen 4 | Stylized figurine for 3D conversion | | "architectural interior, modern kitchen, natural window light, wide angle" | Flux 2 | Photorealistic interior | | "pixel art game character, knight with sword, 16-bit style, transparent background" | Flux 2 | Game sprite | ### Model-specific advice **Flux 2** - Write prompts like natural sentences, not keyword lists. Flux understands flowing descriptions better than comma-separated tags. Put the most important element first. No negative prompts needed. For photorealism, mention camera models ("shot on Fujifilm X-T5") and lighting setups. **ImageGen 4** - Great at following complex instructions with minimal prompting. Produces photorealistic results with simple descriptions. Good at text rendering in images. **GPT-Image-1** - Excellent at following detailed instructions. Works well with structured prompts (scene, subject, details, constraints). Best for text-in-image and complex compositions. ### For Image to 3D conversion If you're generating an image to convert to a 3D model afterward, add these to your prompt: - "clean white background" or "studio lighting" - makes the AI focus on the object - "full body" or "complete object visible" - prevents cropping - "single object, centered" - avoids confusion about what to convert - "3D render style" or "stylized 3D" - produces images that convert to 3D better than flat illustrations ## Image Editing Prompts When editing images in Image Studio (using GPT-Image-2 Edit, Gemini Edit, or Nano Banana), the key principle is: **say what to change AND what to keep**. ### The formula **What to change + what to preserve + style/quality constraints** ### Prompts that work | Prompt | What it does | | --- | --- | | "turn this person into a LEGO minifigure, plastic material, clean background. Keep the pose and general appearance." | Style transfer while preserving identity | | "remove the background. Keep the subject exactly as is." | Clean background removal | | "change the lighting to warm sunset. Keep everything else the same." | Relighting | | "turn into a stylized 3D figurine, full body, clean background" | Convert photo to 3D-ready reference | | "make it Ghibli style. Preserve the composition and subject position." | Art style transfer | | "replace the jacket with a navy windbreaker. Do not change face, pose, lighting, or background." | Surgical clothing swap | ### Key principles 1. **Be specific about what stays the same** - "keep the face, pose, lighting, and background unchanged" prevents drift 2. **One change at a time** - don't try to change background, clothing, and style in one prompt. Do it in rounds. 3. **Describe the target, not the process** - "warm sunset lighting" works better than "make the image warmer" 4. **For text in images** - put the exact text in quotes: 'write "SALE 50% OFF" in bold white letters, top center' 5. **Match lighting** - if the original has warm light, don't ask for cold light in the edit or the result will look composited ### Model comparison for editing | Model | Best for | Notes | | --- | --- | --- | | GPT-Image-2 Edit | Precise edits, text rendering, photorealism | Most capable overall, best at following complex instructions | | Gemini Edit | Quick style changes, creative edits | Fast, good for experimentation | | Nano Banana 2 | General editing, background changes | Good balance of speed and quality | ## The Best Workflow The highest quality results usually come from combining image generation with 3D generation: 1. **Generate a reference image** in Image Studio with a clean prompt 2. **Edit it** if needed (remove background, adjust style, fix details) 3. **Convert to 3D** using Image to 3D with Prism 3.1 This gives you more control than Text to 3D alone, because you can see and refine the 2D reference before committing to 3D generation. [Try Text to 3D](https://www.3daistudio.com/TextTo3D) - [Open Image Studio](https://www.3daistudio.com/ImageGeneration) - [Try Image to 3D](https://www.3daistudio.com/ImageTo3D) ## Frequently Asked Questions ### What's the ideal prompt length for AI 3D generation? For Text to 3D in 3D AI Studio, 5-20 words works best. Anything shorter than 5 words is usually too vague, and anything over 20 words tends to confuse the model. Focus on naming the object, its key material, and the style you want. A prompt like "medieval stone castle with tall tower" is more effective than a 50-word paragraph. ### Do different AI models need different prompting styles? Yes. In 3D AI Studio, Tripo P1 works best with short, punchy prompts (5-10 words), while Prism 3.1 and Hunyuan 3.1 Pro handle more detailed descriptions. For image generation, Flux 2 prefers natural sentences over comma-separated keywords, and ImageGen 4 follows complex instructions well with minimal prompting. Check the model-specific tips table on this page for guidance. ### How do I write prompts for game-ready 3D assets? Add keywords like "game ready," "game asset," or "clean topology" to your prompt in 3D AI Studio. Specify the art style ("low-poly," "stylized," "realistic PBR") and mention "one solid piece" to prevent floating geometry. For the Tripo P1 model specifically, just adding "game ready" to any object prompt optimizes the output for game engines. ### Why does my AI 3D model have floating parts or broken geometry? This usually happens when the prompt describes multiple objects or a complex scene. In 3D AI Studio, always describe a single object and add "one solid piece" or "single connected mesh" to your prompt. Avoid prompts like "a sword and shield" - generate each item separately instead. ### What prompt keywords improve texture quality? Adding "PBR texture," "detailed surface," or specific material names ("brushed steel," "worn leather," "polished marble") in 3D AI Studio gives the AI clearer direction for textures. Also mention "realistic" or "photorealistic" if you want lifelike surface detail. Enable the PBR toggle in settings to get normal, roughness, and metallic maps alongside the base color. ### How do I prompt for images that convert well to 3D? Include "clean white background," "single object, centered," and "3D render style" in your image generation prompt. These keywords tell 3D AI Studio's Image Studio to produce isolated objects with good depth cues that the Image to 3D converter handles easily. Avoid flat illustration styles, and use "full body" for characters to prevent cropping. --- # Texture Generator Source: https://docs.3daistudio.com/textures/texture-generator The [AI texture generator](https://www.3daistudio.com/TextureGenerator) textures an entire 3D model with AI. Upload a GLB, FBX or OBJ, describe the look or add a reference image, and get back a fully textured GLB with optional PBR maps in 1-3 minutes for 20-35 credits. | | | | --- | --- | | **Cost** | 20-35 credits, depending on engine and quality | | **Time** | 1-3 minutes | | **Input** | GLB, FBX or OBJ up to 120MB (200MB with Hunyuan) | | **Guidance** | Text prompt, reference image, or four reference views (Prism) | | **Output** | Textured GLB, optional PBR maps (base color, roughness, metalness, normal) | | **Geometry** | Unchanged, only the surface is replaced | ![Texture Generator model upload interface](https://docs.3daistudio.com/assets/images/docs/tex-gen-input.jpg) *Upload your 3D model and choose an AI texturing engine on the right sidebar.* ## How do I texture a 3D model with AI? Upload the model, pick an engine, describe the material or add a reference image, and click Generate. The AI paints new textures onto the model's UV layout and returns a textured GLB. 1. Upload your 3D model (GLB, FBX or OBJ, up to 120MB) or select one from your dashboard 2. Pick an AI engine. Prism is the default and the best choice for most models 3. Enter a prompt such as *weathered wooden surface with visible grain*, or upload a reference image 4. Adjust quality and material settings, and enable PBR for game engines 5. Click **Generate** (or press Ctrl+Enter) 6. Preview the result and download the textured GLB ![Texture generation in progress](https://docs.3daistudio.com/assets/images/docs/tex-gen-generating.jpg) *The AI analyzes your model and generates new textures. This takes 1-3 minutes depending on the engine.* ## Which texturing engine should I use? Start with **Prism**: it handles the widest range of styles and produces the most consistent results. Use **Forge** when you have a reference image to match, **Meshy** when the existing UV layout must be kept, and **Hunyuan** for files between 120MB and 200MB. | Engine | Credits | Best for | | --- | --- | --- | | **Prism** (default) | 20 (fast or standard), 30 (HD), 35 (max) | Best overall quality, 4K and 8K support | | **Forge** | 20 | Texturing from a reference image | | **Meshy** | 25 | Clean UV handling, keeps the original UV layout | | **Hunyuan** | 20 | PBR materials, files up to 200MB | ### Prism settings Prism's Texture Model setting picks the texturing pass: Latest (experimental, the default here) or Standard. Prism 2.5 stays available for reproducing earlier results. - **Texture quality:** Fast (same price as Standard, quicker, less detail, Latest texture model only), Standard, HD (30 credits, sharper 4K textures) or Max (35 credits, highest resolution) - **Texture prompt:** describe the look, upload a reference image, or provide all four views. Leave it empty on a model from your dashboard and the generator re-textures it from that model's original reference image or prompt. Uploaded models need a prompt or reference image. - **Neutral Lighting:** on by default with the Latest texture model. It removes lighting baked into the reference image before texturing. Turn it off to keep the original shading. - **PBR:** generates roughness, metalness and normal maps alongside the base color - **Texture alignment:** "original image" matches the model's existing look, "geometry" ignores existing textures and works from the shape alone - **Texture seed:** lock a seed to get repeatable results ### Forge settings Forge needs a reference image showing the look you want and matches its style and materials. - **Reference image:** required - **Material type:** PBR (physically based) or Shaded (simpler, baked lighting) - **Resolution:** Basic or High ### Meshy settings Meshy focuses on clean UV handling and consistent surface coverage. - **AI model:** Meshy-7 (newest, recommended), Meshy-6 (adds lighting removal) or Meshy-6 Lite (budget, 2K textures only) - **Preserve original UV:** keeps the model's existing UV layout - **PBR:** generates full PBR material maps ### Hunyuan settings Hunyuan is Tencent's texturing model. It accepts a text prompt or a reference image, but not both at once. - **Input mode:** text prompt or image prompt - **PBR:** generates PBR material maps - **File size limit:** 200MB ## Which PBR maps do I get? With PBR enabled, the GLB contains four maps in a standard glTF material: base color (albedo), roughness, metalness and normal. Unity, Unreal Engine, Blender and any glTF-compatible renderer read them directly, with no manual map assignment. ![Generated texture result with PBR materials applied](https://docs.3daistudio.com/assets/images/docs/tex-gen-result.jpg) *The finished result with AI-generated textures applied. You can toggle between the original and textured version.* ## Does my model need UVs before texturing? Not strictly, but models with clean UVs produce better results with every engine. If the model has no UVs or a messy layout, run [UV Unfold](https://docs.3daistudio.com/processing/uv-unfold) first. It costs 5-15 credits depending on the engine. ## What can I do after generating? - **Download** the textured GLB - **Refine areas by hand** with "Paint in 3D", which opens the result in the [Texture Painter](https://docs.3daistudio.com/textures/texture-painter) - **Save to your dashboard** to keep it in your project library - **Convert** to FBX, OBJ or other formats with [Format Conversion](https://docs.3daistudio.com/utilities/format-conversion) > [!NOTE] > If the AI misses a section of the model, switch Prism's texture alignment to "geometry" and generate again. Adding a reference image next to the prompt gives more control over the final look. ## Related tools - Need a model first? Generate one with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) or [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d) - Paint textures manually with the [Texture Painter](https://docs.3daistudio.com/textures/texture-painter) - Create tileable materials with [Material AI](https://docs.3daistudio.com/textures/material-ai) and [Seamless Texture AI](https://docs.3daistudio.com/textures/seamless-texture) - Pick the right download format in [Export Formats](https://docs.3daistudio.com/export-formats) ## Frequently Asked Questions ### Does retexturing change the shape of my model? No. The Texture Generator only replaces the surface textures. The geometry, vertex positions and rigging stay exactly as they were. ### What is the difference between Texture Generator and Texture Painter? The Texture Generator retextures the whole model in one pass from a prompt or reference image. The Texture Painter gives manual control with brushes, layers and angle-by-angle AI projection. Use the Generator for fast full-model results and the Painter for precise work on specific areas. --- # Texture Painter Source: https://docs.3daistudio.com/textures/texture-painter The Texture Painter is an in-browser workspace for painting directly on 3D models. It combines manual brushwork with AI-powered texture projections and a full layer system. Paint at up to 8K resolution and export with PBR materials. This page covers V2, the current beta version. ![Texture Painter full workspace with 3D model and tools](https://docs.3daistudio.com/assets/images/docs/tex-painter-example.jpg) *The Texture Painter workspace. Left: 3D painting viewport. Right: AI preview and generation panel.* ## Tools | Tool | What it does | |------|-------------| | Color Brush | Freehand painting directly on the 3D surface | | Bucket Fill | Flood-fill connected UV regions with a color | | Eraser | Remove paint from any layer | | Image Brush | Stamp images from the built-in library onto the surface | | Eyedropper | Pick a color from anywhere on the model | ## Layer Types The Texture Painter uses a layer system similar to Photoshop. You can stack, reorder, and blend layers to build up complex textures. - **Paint layers** - Manual brush painting. Each stroke goes onto the active paint layer. - **Project layers** - AI-generated textures projected from a camera capture onto the model's UV map. - **Decal layers** - An AI-generated or uploaded image projected locally onto a specific area. Supports zoom and rotation controls for precise placement. ![Layer panel showing paint and projection layers](https://docs.3daistudio.com/assets/images/docs/tex-painter-layer.jpg) *The Apply Layer dialog lets you adjust coverage and edge softness before projecting the AI-generated texture onto your model. On the right: the prompt input and AI model selector (GPT Image 2 shown).* ## AI Models for Projection When creating a project layer or decal, you choose an AI model to generate the texture. | Model | Credits | Best for | |-------|---------|----------| | Nano Banana 2 | 4-12 | Flexible resolution, good default choice | | Nano Banana Pro | 8-14 | Highest quality projections | | GPT Image 2 Edit | 3-12 | Precise edits that respect existing content | | Qwen 2 Pro | 5 | Alternative option, flat cost | | ImageGen 4 (decals) | 4 | Generate decal images from a text prompt | ## How AI Projection Works AI projection is the core workflow for texturing a model quickly. Here is the process: 1. Position your model in the 3D viewer at the angle you want to texture 2. Capture the current camera view 3. Write a prompt describing the texture you want (for example, "worn leather with brass buckles") 4. The AI generates a texture based on that view and your prompt 5. The generated texture is projected back onto the model's UV map 6. Rotate the model to a new angle and repeat for full coverage You build up texture coverage by projecting from multiple angles. Each projection becomes its own layer, so you can adjust opacity or erase sections that overlap poorly. ![Model mid-painting with partial texture coverage](https://docs.3daistudio.com/assets/images/docs/tex-painter-painted.jpg) *A model being textured with a combination of AI projection and manual brush painting.* ## Multiview Mode Instead of manually capturing one angle at a time, multiview mode captures multiple angles automatically. This gives more consistent texturing across the full model surface since the AI sees the model from several directions at once. ## Resolution - **Paint canvas:** 1K to 8K resolution - **Export:** Up to 12K Higher resolution means more detail but slower performance. 4K is a good balance for most projects. ## Export Options - **GLB** - Standard textured model - **GLB with PBR materials** - Includes roughness, metalness, and normal maps - **Optimized GLB** - Compressed for web and real-time use - **Screenshots** - Export the current viewport as an image ## Keyboard Shortcuts | Shortcut | Action | |----------|--------| | Ctrl+Z | Undo | | Ctrl+Y | Redo | | Z | Cycle view modes (textured / solid / unlit / normals) | | B | Toggle sidebar | ## Tips > [!NOTE] > Start with AI projection to lay down base textures, then switch to manual painting for refinement and detail work. - Use the "Preserve composition" toggle when you want to add details without changing existing textures - Work at 4K resolution for a good balance of quality and performance - Export with PBR materials if your target supports it (Unity, Unreal, Blender) - Use decal layers for logos, labels, and small detail patches that need precise placement - The eraser works per-layer, so you can clean up projection overlap without affecting other layers ## Related Tools - Need a model first? Generate one with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) or [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d) - Want AI-generated textures instead of painting? Use the [AI texture generator](https://www.3daistudio.com/TextureGenerator) ([guide](https://docs.3daistudio.com/textures/texture-generator)) - Need clean UVs before painting? Run [UV Unfold](https://docs.3daistudio.com/processing/uv-unfold) first - Need to reduce poly count before texturing? Use the [Remesh Tool](https://docs.3daistudio.com/processing/remesh) - Importing the painted model into a game engine? Check [Integrations](https://docs.3daistudio.com/integrations) for format tips ## Frequently Asked Questions ### Can I paint textures on a 3D model without downloading software? Yes. The Texture Painter in 3D AI Studio runs entirely in your browser. There is nothing to install. You get brush tools, layers, AI projection, and up to 8K resolution, all from a web interface that works on any modern computer. ### How is this different from Substance Painter? The 3D AI Studio Texture Painter is browser-based and combines manual painting with AI-generated texture projections. You do not need a license or local install. It is simpler for quick texturing jobs, though Substance Painter still offers more advanced features like particle brushes and channel painting for AAA production pipelines. ### How does AI texture projection work? You position your model at an angle, capture the view, and write a prompt describing what you want. The AI in 3D AI Studio generates a texture for that viewpoint and projects it onto the model's UV map. You repeat from different angles to build up full coverage. Each projection becomes a separate layer you can adjust or erase. ### What resolution can I paint at? The painting canvas in 3D AI Studio supports 1K up to 8K resolution, and exports go up to 12K. For most projects, 4K gives a good balance between detail and browser performance. Higher resolutions work best on machines with dedicated GPUs. ### How many credits does texture painting cost? The manual brush tools are free to use. AI projection costs vary by model: Nano Banana 2 uses 4 to 12 credits, GPT Image 2 Edit uses 3 to 12, and Nano Banana Pro uses 8 to 14. Each projection is one credit charge, so a typical model painted from 4 to 6 angles might cost 20 to 50 credits total in 3D AI Studio. ### Can I export the painted model with PBR materials? Yes. 3D AI Studio lets you export your painted model as a GLB with full PBR materials, including roughness, metalness, and normal maps. There is also an optimized GLB option with compression for web and real-time applications. --- # AI Material Generator Source: https://docs.3daistudio.com/textures/material-ai The AI Material Generator creates full PBR material map sets from text, photos, or existing images. You get base color, normal, roughness, metalness, and height maps, all ready for use in any 3D engine or design tool. ![Generating a material from a text description](https://docs.3daistudio.com/assets/images/docs/tex-material-generating.jpg) *Generating a PBR material set from a text description. Each map is created individually.* ## Three Modes | Mode | What it does | Credits | |------|-------------|---------| | Describe Material | Type a text description and get a full PBR material set | 10 | | Extract from Photo | Upload a photo, describe a region, and extract that surface as a material | 15 | | Image to Maps | Upload a texture image and get PBR maps generated from it | 7 | ![Complete material result with all PBR maps](https://docs.3daistudio.com/assets/images/docs/tex-material-result.jpg) *A generated brick wall material with all PBR maps: base color, normal, roughness, metalness, and height.* ## PBR Maps Generated Every generation produces up to five maps: - **Base Color** - the diffuse color of the surface - **Normal** - surface detail that affects lighting without changing geometry - **Roughness** - controls how matte or glossy the surface appears - **Metalness** - defines which areas are metallic vs non-metallic - **Height** - depth information for displacement and parallax effects Each map can be downloaded individually. ## Settings - **Tiling direction** - Both, Horizontal, or Vertical - **Output format** - PNG, JPEG, or WebP - **Per-map toggles** - enable or disable individual maps - **Seed** - lock in a value for reproducible results - **Upscale** - higher resolution output (+40 credits, paid plans only) - **Material presets** - starting points like Brushed Steel, Worn Leather, Mossy Stone, and more ## Preview Tabs The tool has three preview tabs so you can inspect results before downloading: 1. **Material Preview** - a 3D sphere with all maps applied, showing how the material looks on a surface 2. **All Maps** - a grid view of each individual map side by side 3. **Tiling Check** - a seamless repeat preview to verify the material tiles cleanly ![Extracting a material from a photo](https://docs.3daistudio.com/assets/images/docs/tex-material-extract.jpg) *Extract mode: upload a photo and describe the region you want to turn into a tileable material.* ## Tips - Use **Describe Material** when you want to create an original material from your imagination. Just type what you want and the AI builds it. - Use **Extract from Photo** when you have a real-world surface you want to replicate. Upload a photo of brick, wood, fabric, or any surface and describe which part to extract. - Use **Image to Maps** when you already have a base color texture and need the other PBR maps (normal, roughness, metalness, height) generated from it. - Material presets are a great starting point. Pick one that is close to what you need, then customize the description to get exactly what you want. ## Access Free users can generate materials and preview results. Downloads are locked to paid plans. ## Related Tools - Apply generated materials to 3D models using the [Texture Painter](https://docs.3daistudio.com/textures/texture-painter) or [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) - Need a tileable texture instead? Use the [Seamless Texture Generator](https://docs.3daistudio.com/textures/seamless-texture) - Need PBR maps from an existing image? Try [PBR Extract](https://docs.3daistudio.com/textures/pbr-extract) - Generate 3D models to apply materials to with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) or [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d) ## Frequently Asked Questions ### Is the AI Material Generator free to use? Free users can generate materials and preview all PBR maps in 3D AI Studio. Downloading the map files requires a paid plan. Generation costs range from 7 to 15 credits depending on the mode, with "Describe Material" at 10 credits and "Extract from Photo" at 15. ### What PBR maps does it generate? The Material Generator in 3D AI Studio produces five maps: base color, normal, roughness, metalness, and height. You can toggle individual maps on or off before generating, and each map downloads as a separate image file in PNG, JPEG, or WebP format. ### Can I create a material from a real-world photo? Yes. Use the "Extract from Photo" mode in 3D AI Studio to upload a photo of any surface, like brick, wood, fabric, or metal. Describe which region you want extracted, and the AI turns it into a tileable PBR material set with all five maps. ### How do I use generated materials in Blender or Unity? Download the individual map files and plug them into your material shader. In Blender, connect base color to the Principled BSDF color input, normal to Normal Map node, roughness to Roughness, metalness to Metallic, and height to a Displacement node. Materials generated in 3D AI Studio follow standard PBR conventions, so they work in any engine. ### What is the difference between Material Generator and Seamless Texture Generator? Both tools in 3D AI Studio produce PBR map sets. The Material Generator has three modes including photo extraction and is more general purpose. The Seamless Texture Generator is specifically optimized for tiling quality and includes a repeat preview. If you need textures for large repeating surfaces like floors or walls, Seamless is the better choice. ### Can I upscale the generated materials to higher resolution? Yes. 3D AI Studio offers an upscale option that produces higher resolution maps for an additional 40 credits. This feature is available on paid plans only and works with any of the three generation modes. --- # Seamless Texture Generator Source: https://docs.3daistudio.com/textures/seamless-texture The Seamless Texture Generator creates tileable PBR texture sets from text descriptions or images. Every map it produces tiles without visible seams, making it ideal for surfaces that repeat across large areas. ![Seamless texture result with tiling preview](https://docs.3daistudio.com/assets/images/docs/tex-seamless-result.jpg) *A generated rock floor texture shown in the seamless preview with tiling check. The maps tile without visible seams.* ## Two Modes | Mode | Credits | |------|---------| | Text to Material | 10 | | Image to Material | 7 | There is no "Extract from Photo" mode here. If you need to extract a material from a photo, use the [Material Generator](https://docs.3daistudio.com/textures/material-ai) instead. ## PBR Maps Every generation produces a full set of maps that all tile seamlessly: - **Base Color** - the diffuse surface color - **Normal** - surface detail for lighting - **Roughness** - matte vs glossy areas - **Metalness** - metallic vs non-metallic regions - **Height** - depth for displacement ## Preview Tabs 1. **Seamless Preview** - a tiled repeat check with adjustable tile count, so you can verify the texture repeats cleanly at different scales 2. **3D Preview** - a textured mesh with adjustable strength sliders for normal, height, roughness, and metalness maps 3. **All Maps** - a grid view of each individual map ## Settings - **Tiling direction** - Both, Horizontal, or Vertical - **Output format** - PNG, JPEG, or WebP - **Seed** - for reproducible results - **Strength slider** - controls how much the generation follows your input - **Map strength sliders** - adjust individual map intensities in the 3D preview - **Per-map toggles** - enable or disable specific maps - **Upscale** - higher resolution output (+40 credits, paid plans only) ## Difference from Material Generator Both tools generate PBR material sets, but they have different strengths: - **Seamless Texture Generator** is specifically optimized for tileable results and includes a seamless tiling preview tab. Best when you know you need a repeating surface. - **Material Generator** has an additional "Extract from Photo" mode and is more general purpose. Both use the same AI backend, but Seamless focuses on tiling quality. ## Use Cases - Game environments: floors, walls, terrain, ceilings - Architectural visualization: building materials, interior surfaces - Product design: surface finishes, packaging textures - 3D scene backgrounds: natural surfaces, patterns ## Access Free users can generate textures and preview results. Downloads are locked to paid plans. ## Related Tools - Need full PBR material sets? Use the [Material Generator](https://docs.3daistudio.com/textures/material-ai) - Need PBR maps from an existing texture? Try [PBR Extract](https://docs.3daistudio.com/textures/pbr-extract) - Apply textures to 3D models with the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) or [Texture Painter](https://docs.3daistudio.com/textures/texture-painter) - Building game environments? See [Export Formats](https://docs.3daistudio.com/export-formats) for format recommendations and [Integrations](https://docs.3daistudio.com/integrations) for engine-specific import guides ## Frequently Asked Questions ### How do I generate seamless textures from a text description? Open the Seamless Texture Generator in 3D AI Studio and select "Text to Material" mode. Type a description of the surface you want, like "mossy cobblestone path" or "brushed aluminum panel." The AI generates a full PBR map set that tiles perfectly with no visible seams. It costs 10 credits per generation. ### Are the generated textures actually tileable? Yes. The Seamless Texture Generator in 3D AI Studio is specifically optimized for tiling. Use the built-in Seamless Preview tab to verify the texture repeats cleanly at different tile counts. The AI handles edge matching internally, so you do not need to fix seams manually. ### Can I use these textures for game development? Absolutely. The PBR maps from 3D AI Studio (base color, normal, roughness, metalness, height) work directly in Unity, Unreal Engine, Godot, and any other engine that supports standard PBR materials. The textures are designed for repeating surfaces like floors, walls, and terrain that are common in game environments. ### What is the difference between "Text to Material" and "Image to Material" modes? "Text to Material" creates an original texture from your written description. "Image to Material" takes an existing image you upload and generates a seamless, tileable PBR map set from it. Both modes in 3D AI Studio produce the same five maps, but Image mode costs only 7 credits compared to 10 for Text mode. ### Can I control the tiling direction? Yes. 3D AI Studio lets you set tiling direction to Both, Horizontal, or Vertical. Use "Both" for surfaces that repeat in all directions like floors. Use Horizontal or Vertical for surfaces that only repeat along one axis, like a fence or a baseboard. ### How is this different from downloading textures from a texture library? Unlike static texture libraries, the Seamless Texture Generator in 3D AI Studio creates original textures on demand from any description you can imagine. You are not limited to what someone has already photographed or created. You can also fine-tune results with seed values and strength controls for exactly the look you need. --- # PBR Map Generator Source: https://docs.3daistudio.com/textures/pbr-extract The PBR Map Generator extracts a single PBR map type from any uploaded photo or texture image. Upload your image, pick the map type you need, and the AI generates it. ![PBR Map Generator with before/after comparison](https://docs.3daistudio.com/assets/images/docs/tex-pbr-extract.jpg) *A normal map generated from a source image. Use the before/after slider to compare the original with the extracted map.* ## Map Types | Map | What it generates | |-----|------------------| | Normal | Surface detail for realistic lighting | | Roughness | Matte vs glossy areas | | Height / Displacement | Depth for parallax and displacement mapping | | Ambient Occlusion | Crevice and cavity shadows | | Metallic | Metal vs non-metal regions | | Depth | Foreground/background distance map | ## AI Models | Model | Credits | Resolution | |-------|---------|-----------| | Nano Banana 2 | 4 | 1K | | Gemini Edit | 6 | 1K | | Nano Banana Pro | 8-16 | 1K, 2K, 4K | Higher resolution options cost more credits. 4K resolution is available on paid plans only. ## How to Use 1. Upload a photo or texture image 2. Select the map type you want (normal, roughness, height, etc.) 3. Pick an AI model based on your quality and resolution needs 4. Click generate 5. Preview the result with the before/after slider 6. Download the map ## Difference from Material Generator These two tools serve different purposes: - **PBR Map Generator** extracts ONE map at a time from a photo or existing texture. Use it when you need a specific map from an image you already have. - **Material Generator** creates a FULL SET of maps (base color, normal, roughness, metalness, height) from scratch or from a description. Use it when you need a complete material. If you have a texture and just need a normal map from it, use PBR Map Generator. If you want to create an entire material from a text prompt, use Material Generator. ## Access Free users can generate maps using their credits. Downloads are locked to paid plans. 4K resolution requires a paid plan. ## Related Tools - Need a complete material set from scratch? Use the [Material Generator](https://docs.3daistudio.com/textures/material-ai) instead - Want tileable textures for game environments? Try the [Seamless Texture Generator](https://docs.3daistudio.com/textures/seamless-texture) - Apply extracted maps to a 3D model with the [Texture Painter](https://docs.3daistudio.com/textures/texture-painter) - Generate 3D models to texture with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) or [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d) ## Frequently Asked Questions ### Can I generate a normal map from a regular photo? Yes. Upload any photo or texture image to the PBR Map Generator in 3D AI Studio, select "Normal" as the map type, and the AI extracts surface detail information into a proper normal map. It works with photos of real surfaces, hand-painted textures, or any flat image. ### What resolution are the extracted PBR maps? Resolution depends on the AI model you choose in 3D AI Studio. Nano Banana 2 and Gemini Edit output at 1K resolution. Nano Banana Pro offers 1K, 2K, or 4K, with higher resolutions costing more credits. 4K resolution is available on paid plans only. ### How many credits does it cost to extract a PBR map? A single map extraction in 3D AI Studio costs 4 to 16 credits depending on the AI model and resolution. Nano Banana 2 is the cheapest at 4 credits for 1K. Nano Banana Pro at 4K is 16 credits. If you need all five PBR maps from one image, consider using the Material Generator instead, which creates the full set for 7 credits. ### Can I extract a roughness map from a texture I already have? Yes. The PBR Map Generator in 3D AI Studio is designed exactly for this. Upload your existing base color texture, select "Roughness" as the map type, and the AI analyzes the surface to determine which areas should be glossy and which should be matte. The output is a grayscale roughness map ready for your shader. ### What is the difference between height maps and depth maps? In 3D AI Studio, the Height map type generates displacement data showing surface depth, useful for parallax mapping and tessellation in game engines. The Depth map type generates foreground-to-background distance information, more useful for compositing and post-processing effects. For texturing workflows, you typically want the Height map. ### Should I use PBR Extract or the Material Generator? Use the PBR Map Generator in 3D AI Studio when you already have a texture and just need one specific map from it, like extracting a normal map from your base color. Use the Material Generator when you want to create a complete set of all five PBR maps from scratch or from a description. --- # Remesh / Retopology Source: https://docs.3daistudio.com/processing/remesh The Remesh tool rebuilds a model's topology to reduce its polygon count, switch between quads and triangles, and export it in a new format. Textures are baked onto the new mesh, and the cheapest providers cost 0-3 credits. | | | | --- | --- | | **Cost** | Free (Base), 1-3 credits (Prism), 10-15 (Hunyuan), 15 (Pro) | | **Face count** | 1,000 to 200,000 faces (Prism) | | **Topology** | Quads or triangles | | **Texture baking** | 512 to 4,000 pixels, PNG or JPEG | | **Export** | GLB, FBX, OBJ, STL, USDZ, 3MF, BLEND | | **Where** | [Standalone Remesh tool](https://www.3daistudio.com/Tools/Remesh) or Model Actions in your dashboard | ## How do I reduce the polygon count of a 3D model? Open the model in the Remesh tool, choose a provider, set a target face count and click **Remesh**. Turn on texture baking so the original textures are projected onto the lighter mesh. 1. Open the [Remesh tool](https://www.3daistudio.com/Tools/Remesh) and upload a GLB, FBX or OBJ, or open a model from your Dashboard and choose it in **Model Actions** 2. Choose a provider (see below). Prism suits most game and animation work 3. Set the **face count** for your platform and choose **quads** or **triangles** 4. Enable **texture baking** and pick a resolution 5. Pick the output format and click **Remesh** 6. Download the optimized model ## Which remesh provider should I use? Use **Prism** for most models: it gives clean quad or triangle topology for 1-3 credits. Use **Base** when you only need a quick, free reduction or format change, **Hunyuan** for AI-placed topology, and **Pro (Meshy)** for high-poly production assets. | Provider | Credits | Best for | | --- | --- | --- | | **Base** | Free | Format conversion and basic polygon reduction | | **Prism** | 1 (triangles) or 3 (quads) | Clean topology for games and animation | | **Hunyuan** | 10-15 | AI retopology that adapts to the shape | | **Pro (Meshy)** | 15 | High-poly models that must keep fine detail, many output formats | **Base** supports face count targets and triangulated output at no cost, which makes it a good first step when you just need a lighter model. **Prism** offers face counts from 1,000 to 200,000, texture baking, and pre-processing such as vertex cleanup and structure repair. ![Prism remesh settings panel](https://docs.3daistudio.com/assets/images/docs/proc-remesh-prism.jpg) *The Prism provider settings, showing face count, topology type, and texture baking options.* **Hunyuan** decides where edges go based on the model's shape. It puts more polygons on curved and detailed areas and fewer on flat surfaces, for 10-15 credits depending on complexity. **Pro (Meshy)** handles high-polycount models, keeps fine detail while producing proper topology, and outputs many formats, for 15 credits. ## How many polygons should I target? Match the face count to the platform: 1,000-5,000 faces for mobile and web, 5,000-30,000 for PC games, and up to 200,000 for film, renders or 3D printing. | Use | Target face count | | --- | --- | | Mobile games, web viewers | 1,000-5,000 | | PC and console games | 5,000-30,000 | | Film, renders, 3D printing | 10,000-200,000 | For very high-poly source models (over 100,000 faces), turn on **initial decimation** in pre-processing. It gives the remesher a cleaner starting point. ## Should I use quads or triangles? Use quads for models you will animate or subdivide, and triangles for real-time rendering in game engines. With Prism, quads cost 3 credits and triangles 1 credit. | | Quads | Triangles | | --- | --- | --- | | Deformation in animation | Clean | Can crease | | Subdivision surfaces | Works well | Not suitable | | Game engine rendering | Converted to triangles on import | Native | | Prism cost | 3 credits | 1 credit | ## Does remeshing keep my textures? Only with texture baking turned on. Remeshing creates new geometry with a new UV layout, so the original textures no longer line up unless they are baked onto the new mesh. Baking supports 512 to 4,000 pixel textures in PNG or JPEG. ## Which formats can I export after remeshing? Remeshed models download as **GLB** (web and game engines), **FBX** (Unity, Unreal, Maya), **OBJ** (universal), **STL** (3D printing), **USDZ** (Apple AR), **3MF** (color printing) or **BLEND** (Blender). For 3D printing, choose STL or 3MF and turn on structure repair in pre-processing. See [Export Formats](https://docs.3daistudio.com/export-formats) for details. ## Before and after A model before and after remeshing with the Hunyuan provider, viewed in Blender: ![Model before remeshing in Blender](https://docs.3daistudio.com/assets/images/docs/proc-remesh-before.jpg) *Before: the original mesh with uneven topology and high poly count.* ![Model after Hunyuan remesh in Blender](https://docs.3daistudio.com/assets/images/docs/proc-remesh-after.jpg) *After: clean, optimized topology from the Hunyuan remesh. Notice the even polygon distribution.* ## Video walkthrough How Remesh fits into a full game development workflow: [Video: Game Dev Workflow with Remesh](https://www.youtube.com/watch?v=L5emNlNYVfY) ## What's next After remeshing, [add textures](https://docs.3daistudio.com/textures/texture-generator), [rig the model](https://docs.3daistudio.com/processing/rigging) for animation or [split it into parts](https://docs.3daistudio.com/processing/segmentation). For AI retopology into very light low-poly meshes, see [Smart Low Poly](https://docs.3daistudio.com/processing/smart-low-poly). ## Frequently Asked Questions ### Can I convert file formats without changing the mesh? Yes. The free Base provider re-exports the model between GLB, FBX, OBJ, STL, USDZ, 3MF and BLEND without rebuilding the topology. The [Change Format](https://docs.3daistudio.com/utilities/format-conversion) model action does the same for 1 credit. ### What is AI retopology and when should I use it? AI retopology decides where to place edges based on the model's shape, with more detail on curved or complex areas and fewer polygons on flat surfaces. The Hunyuan provider offers it for 10-15 credits. Use it when uniform decimation leaves too little detail where it matters. ### How do I optimize a 3D model for a game engine? Use the Prism provider with triangles, set 1,000-5,000 faces for mobile or 5,000-30,000 for PC, turn on texture baking, and export as FBX for Unity and Unreal Engine or GLB for web and Godot. --- # Smart Low Poly Source: https://docs.3daistudio.com/processing/smart-low-poly Smart Low Poly rebuilds a dense model as a clean, lightweight mesh. Instead of just deleting polygons, it retopologizes the model: edges follow the shape, detail is kept where it matters, and the original textures are baked onto the new mesh. The result is ready for games, AR/VR, the web, and any real-time use. Open any model from your [Dashboard](https://www.3daistudio.com/Dashboard) and choose **Smart Low Poly** in Model Actions. ## Model versions | Version | Credits | Best for | |---------|---------|----------| | **3.0** (newest, default) | 30 (triangles) / 35 (quads) | The best retopology. Also unwraps clean UVs in the same step | | **2.0** | 25 | Clean, edge-preserving retopology | | **1.0** | 15 | Plain, fast decimation to an exact polygon count | Credits are refunded automatically if a conversion fails. ## Settings - **Topology** -- *Triangles* gives a GLB for game engines and real-time viewers. *Quads* gives an FBX with four-sided faces for animation and DCC tools. - **Polygon count** -- *Auto* chooses the best count for the model. *Custom* sets a target: - 3.0 and 2.0: 500-20,000 triangles or 500-10,000 quads - 1.0: required, up to 2,000,000 triangles or 150,000 quads - **Bake textures** (3.0 and 2.0) -- projects the original textures onto the new mesh. Turn it off for an untextured mesh. - **Unwrap UVs** (3.0, off by default, no extra cost) -- gives the new mesh a clean UV layout in the same step. It softens the existing textures, so turn it on when you plan to re-texture or paint the model in the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) or [Texture Painter](https://docs.3daistudio.com/textures/texture-painter). - **Restore original material** (off by default, free) -- keeps the original matte look. Without it, the result can look glossier than the source. Triangle (GLB) results from GLB models only. - **Clean texture seams** (off by default, free) -- removes dark lines along texture seams that can show when the model is viewed from a distance. The file gets slightly larger. Triangle (GLB) results only. ## Limits - **Max file size:** 150 MB - **Accepted formats:** GLB, FBX, OBJ, STL ## Tips - Start with version **3.0** and **Auto** polygon count. Only set a custom count when your target platform has a hard budget. - Choose **Quads** for characters you plan to rig and animate; choose **Triangles** for everything that goes straight into a game engine. - Leave **Unwrap UVs** off for the sharpest textures; turn it on only when you will re-texture or paint the model. - Faces and small details are simplified at 20,000 polygons or fewer. For hero characters, keep a higher-detail version for close-ups. - The low poly texture keeps the resolution of the source texture. Models generated with Fast textures give a soft result; generate with a higher texture quality for sharper results. - Need several detail levels? Use **Add LOD's** in Model Actions instead. ## What's Next With a light mesh in place, you can [rig and animate it](https://docs.3daistudio.com/processing/rigging), [re-texture it](https://docs.3daistudio.com/textures/texture-generator), or [unwrap its UVs](https://docs.3daistudio.com/processing/uv-unfold) with Smart UV. Want a low poly model from the start instead of reducing one? The [AI low poly generator](https://www.3daistudio.com/UseCases/LowPoly) creates low poly models and characters directly. ## Frequently Asked Questions ### What is the difference between Smart Low Poly and Remesh? Smart Low Poly is focused on one job: turning a dense model into a clean, light mesh that keeps its look, with almost no settings. The [Remesh tool](https://docs.3daistudio.com/processing/remesh) offers more providers and export options (format conversion, texture resolution, pre-processing) for production pipelines. ### Which version should I choose? 3.0 is the newest version, gives the best results and is the default. 2.0 is slightly cheaper and still produces clean topology. 1.0 is the cheapest and fastest, but it simply reduces the polygon count without rebuilding the topology, so it needs an exact target polygon count. ### Does Smart Low Poly keep my textures? Yes, with **Bake textures** on (the default), the original textures are projected onto the new mesh. Basic decimation keeps the existing texture layout of the model. ### Why is my quad result an FBX file? GLB cannot store four-sided faces, so quad meshes are delivered as FBX. Choose triangles if you need a GLB. ### Can I get clean UVs at the same time? Yes. With version 3.0, turn on **Unwrap UVs**: the low-poly mesh comes back with a fresh, clean UV layout at no extra cost. It softens the existing textures, so it's best when you plan to re-texture or paint the model. ### Why does my low poly model look glossier or show dark lines on the texture? The conversion can make the material look glossier than the original and leave dark lines along texture seams. Turn on **Restore original material** and **Clean texture seams** (both free) to fix this. ### Why is the texture blurrier than the original? The low poly texture keeps the source's resolution, and the original texture is re-projected onto a much simpler mesh, so some sharpness is lost. Models generated with Fast textures start out softer; generate with a higher texture quality for sharper results. --- # AI Rigging & Animation Source: https://docs.3daistudio.com/processing/rigging The Rigging tool adds a skeleton and skin weights to a 3D character automatically, so it can be animated. Rigging takes about 1-2 minutes and costs 15-20 credits, and a free pre-rig check tells you beforehand whether your model can be rigged. | | | | --- | --- | | **Cost** | 20 credits (Prism) or 15 credits (Meshy), animations 15 credits each | | **Time** | About 1-2 minutes | | **Input** | GLB or FBX character model | | **Body types** | Biped, quadruped, avian, serpentine (Prism); humanoid (Meshy) | | **Output** | GLB or FBX with skeleton and skin weights, Mixamo-compatible (Prism) | | **Animations** | 12 presets with automatic retargeting | ## How do I rig a 3D model with AI? Upload the character, run the free pre-rig check, pick a provider and body type, and click **Rig Model**. The AI places the joints and paints the skin weights for you, which replaces hours of manual work in Blender or Maya. 1. Upload your 3D model (GLB or FBX) in the [Rigging tool](https://www.3daistudio.com/Tools/RiggingTool), or open it from your Dashboard 2. Run the free **pre-rig check** to confirm the model can be rigged 3. Choose a provider and the rig type that matches your character 4. Click **Rig Model** and wait about 1-2 minutes 5. Inspect the result in the rigged model viewer 6. Download the rigged model or apply an animation ![Rigged model showing skeleton overlay](https://docs.3daistudio.com/assets/images/docs/proc-rig-result.jpg) *A rigged character model with the AI-generated skeleton visible. Joints are placed at natural positions for realistic movement.* ## Which rigging provider should I use? Use **Prism** for anything that is not a standard humanoid and when you want a Mixamo-compatible rig. Use **Meshy** for humanoid characters when you want the cheaper option with walk and run animations included. | Provider | Credits | Body types | Extras | | --- | --- | --- | --- | | **Prism** | 20 | Biped, quadruped, avian, serpentine | GLB and FBX export, Mixamo-compatible rig | | **Meshy** | 15 | Humanoid only | Walk and run animations included | ## How do I prepare a model for rigging? Use a single-mesh character in a T-pose or A-pose with clean geometry. Arms held away from the body give the AI a clear separation between limbs and torso, which leads to more accurate joints and skin weights. - **Pose:** T-pose or A-pose works best. Other poses can be rigged, with less precise results for arms close to the body. - **Single mesh:** one connected character rigs better than a model made of many separate parts. - **Textures:** textured models are preferred, because surface details help the AI recognize body parts. - **Clean geometry:** if the topology is messy, run the [Remesh tool](https://docs.3daistudio.com/processing/remesh) first. - **Pre-rig check:** run it before spending credits. If it reports the model as unsuitable, remesh it and check that the limbs are clearly defined. ## How do I animate a rigged model? After rigging, pick one of 12 preset animations. Each costs 15 credits and is retargeted to your character's proportions automatically. | Animation | Animation | Animation | | --- | --- | --- | | Idle | Walk | Run | | Jump | Dance | Wave | | Salute | Boxing | Sword Slash | | Sitting | Yoga | Celebration | ![Animation presets grid](https://docs.3daistudio.com/assets/images/docs/proc-rig-animations.jpg) *The 12 animation presets available after rigging. Click any to preview and apply.* The rigged model viewer lets you check the result before downloading: - **Skeleton overlay** to see where the joints are - **Play, pause and scrub** through the animation - **Speed controls** to slow down or speed up playback - **Lock position** to keep the model centered during movement animations ## Can I use the rig in Mixamo, Unity or Unreal? Yes. Rigged models export as GLB or FBX with the skeleton and skin weights embedded, and Prism rigs follow the Mixamo bone naming convention. That means they work with Mixamo's animation library, Unity's Humanoid system and Unreal Engine's mannequin retargeting. For your own animations, export the rig as FBX and animate it in Blender, Maya or your game engine. The preset animations are optional. Engine-specific import steps are in the [Unity](https://docs.3daistudio.com/integrations/unity), [Unreal Engine](https://docs.3daistudio.com/integrations/unreal) and [Blender](https://docs.3daistudio.com/integrations/blender) guides. ## What's next Add or replace textures with the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator), or manage your rigged models in the [Dashboard](https://www.3daistudio.com/Dashboard). ## Frequently Asked Questions ### How much does AI rigging cost? 15 to 20 credits. Prism costs 20 credits and supports every body type, Meshy costs 15 credits for humanoid characters. Each preset animation applied afterward costs another 15 credits, and the pre-rig check is free. ### Can I rig animals, birds or dragons? Yes, with the Prism provider. Choose quadruped for four-legged animals, avian for birds and winged creatures, and serpentine for snakes and similar bodies. The AI adapts the bone placement to the creature's anatomy. --- # 3D Mesh Segmentation Source: https://docs.3daistudio.com/processing/segmentation The Segmentation tool uses AI to split a single 3D model into its logical parts. Arms, legs, head, clothing, weapons, accessories, whatever makes up your model gets separated into distinct pieces you can work with individually. **Cost:** 15 credits per segmentation ## How It Works 1. Upload a GLB file (single mesh, single material) 2. Click **Segment** to start processing 3. Wait about 90 to 120 seconds 4. View the color-coded result showing each identified part 5. The segmented model saves to your Dashboard That is it. One click after upload. The AI figures out where to split. ![Segmentation settings modal](https://docs.3daistudio.com/assets/images/docs/proc-segment-modal.jpg) *The segmentation tool ready to process an uploaded model.* ![Color-coded segmentation result](https://docs.3daistudio.com/assets/images/docs/proc-segment-result.jpg) *The segmented result with each part shown in a different color. Each colored section is now a separate piece.* ## Requirements - **Input format:** GLB only - **Mesh type:** Single mesh, single material (multi-material models may not segment correctly) - **Processing time:** About 90 to 120 seconds If your model is in a different format, convert it to GLB first using the [Remesh](https://docs.3daistudio.com/processing/remesh) tool with the Base (free) provider. ## Use Cases - **Multi-material characters:** Separate a character into head, body, arms, and legs so you can assign different materials or textures to each part. - **Multi-color 3D printing:** Split a model into parts for printing in different filament colors, then assemble afterward. - **Game LODs and modularity:** Isolate parts for level-of-detail systems, modular character systems, or destructible objects. - **Selective editing:** Extract just the part you want to modify (a helmet, a weapon, a face) without affecting the rest of the model. ## Tips - Works best on characters and objects with clearly distinct parts. A character wearing armor segments better than a smooth abstract sculpture. - If your model has multiple meshes or materials already, merge them into a single mesh first for best results. - After segmentation, you can use [Remesh](https://docs.3daistudio.com/processing/remesh) on individual parts if they need topology cleanup. ## What's Next After segmenting, you can texture individual parts with the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) or export parts in different formats using [Remesh](https://docs.3daistudio.com/processing/remesh). ## Frequently Asked Questions ### How does AI segmentation decide where to split the model? The AI in 3D AI Studio analyzes the geometry and surface features to identify logical boundaries, like where an arm meets a torso or where a helmet sits on a head. It uses shape recognition to find natural separation points. Characters with distinct parts (armor, clothing, accessories) segment more cleanly than smooth organic shapes. ### Can I control how many parts the model gets split into? Currently, the segmentation in 3D AI Studio is fully automatic. The AI determines the number and boundaries of parts based on the model's structure. You cannot specify a target part count. If you need to merge some parts back together afterward, you can do that in Blender or your 3D tool of choice. ### What format does the segmented model come in? The segmented model saves to your 3D AI Studio Dashboard as a GLB file with each part as a separate mesh. From there, you can use the Remesh tool to convert individual parts to FBX, OBJ, STL, or other formats as needed for your workflow. ### Can I use segmentation for multi-color 3D printing? Yes. This is one of the main use cases. Split your model into parts in 3D AI Studio, then print each piece in a different filament color and assemble them afterward. The tool separates parts cleanly enough that they fit back together at the original boundaries. Export each part as STL for your slicer software. ### Why does it require a single-mesh, single-material model? The segmentation AI in 3D AI Studio works by analyzing continuous geometry to find logical split points. If your model already has multiple separate meshes or materials, the algorithm cannot reason about the full shape properly. Merge your model into one mesh first, either in Blender or by running it through the Remesh tool with the free Base provider. ### How long does segmentation take and what does it cost? Processing takes about 90 to 120 seconds in 3D AI Studio and costs 15 credits per segmentation. There is no preview or partial result. The AI processes the full model and delivers all segmented parts at once when complete. --- # UV Unfold Source: https://docs.3daistudio.com/processing/uv-unfold The UV Unfold tool automatically unwraps UV coordinates for your 3D model. Good UV maps are essential for applying textures accurately. This tool handles the unwrapping in one click, saving you the manual work of placing seams and laying out UV islands. ## Engines | Engine | Credits | Max size | Formats | Keeps textures | |--------|---------|----------|---------|----------------| | **Smart UV** | 15 | 80,000 triangles | GLB, FBX, OBJ, STL | Yes (plus rigging and animation) | | **Meshy** | 8 | 40,000 faces | GLB, FBX, OBJ | No | | **Hunyuan** | 5 | 30,000 faces | GLB, GLTF, FBX, OBJ | No | Credits are refunded automatically if an unwrap fails. **Smart UV** re-unwraps the model and then re-bakes its existing textures onto the new layout, so the model looks exactly the same afterwards, just with clean UVs. Rigged and animated models keep their skeleton and animation. Turn off **Keep textures** if you only want geometry with new UVs (for example, because you will re-texture the model anyway). Smart UV also saves a **UV atlas** image to your dashboard: a flat preview of the new UV islands, handy for checking the layout or as a guide for hand-painting. It is free and can be turned off with **Store UV atlas image separately**. The atlas is a preview of the layout, not a texture. **Meshy** and **Hunyuan** generate fresh UVs only. Existing textures do not carry over, so plan to re-texture afterwards. ## How It Works 1. Open a model from your Dashboard and choose **UV Unfold**, or upload a model in the tool 2. Pick an engine 3. Click **Unfold UVs** 4. The AI analyzes the geometry, picks natural seam locations, and generates a clean UV layout 5. The result appears in your Dashboard The AI places seams along natural edges and hidden areas to minimize visible texture stretching. The result is a model ready for texturing in the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator), [Texture Painter](https://docs.3daistudio.com/textures/texture-painter), or external tools like Substance Painter. ![UV Unfold tool interface](https://docs.3daistudio.com/assets/images/docs/proc-uv-unfold.jpg) *The UV Unfold tool showing a model ready for automatic UV unwrapping.* ## Limits - **Smart UV:** up to 80,000 triangles (a quad counts as two) and 150 MB - **Meshy:** up to 40,000 faces - **Hunyuan:** up to 30,000 faces and 120 MB If your model is denser than the engine allows, reduce it first with [Smart Low Poly](https://docs.3daistudio.com/processing/smart-low-poly). Its version 3.0 also unwraps UVs in the same step, so for dense models you often don't need a separate unwrap at all. ## What's Next With clean UVs in place, head to the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) to add materials and surface detail, or use the [Texture Painter](https://docs.3daistudio.com/textures/texture-painter) for hand-directed texture work. ## Frequently Asked Questions ### Why do I need UV unwrapping for AI-generated 3D models? Many AI-generated models come with poor or missing UV coordinates. Without proper UVs, textures stretch, distort, or do not apply at all. Running UV Unfold in 3D AI Studio gives your model a clean UV layout so the Texture Generator and Texture Painter can paint surfaces accurately. ### How much does automatic UV unwrapping cost? Hunyuan costs 5 credits, Meshy 8 credits and Smart UV 15 credits per model. Smart UV costs more because it also carries your existing textures, rigging and animation over to the new layout. ### Does UV Unfold preserve my existing textures? With **Smart UV**, yes: the existing textures are re-baked onto the new UV layout (Keep textures, on by default), and rigging and animation are kept too. Hunyuan and Meshy regenerate the UV coordinates from scratch, so existing textures no longer align and the model needs re-texturing. ### What is the UV atlas image? A flat picture of the new UV islands that Smart UV saves next to your model. It shows how the model's surface is laid out in texture space, which is useful for checking the unwrap or as a template for painting textures by hand. It is not a texture itself, and it is free. ### What is the maximum model size for UV unwrapping? Smart UV accepts up to 80,000 triangles, Meshy up to 40,000 faces and Hunyuan up to 30,000 faces. If your model is denser, run [Smart Low Poly](https://docs.3daistudio.com/processing/smart-low-poly) first, or turn on its Unwrap UVs option (version 3.0) to reduce and unwrap in one step. ### How does the AI decide where to place UV seams? The AI analyzes the geometry and places seams along natural edges, creases, and hidden areas like the back of limbs or underside of objects. This minimizes visible texture stretching on prominent surfaces while keeping seam artifacts in less noticeable locations. ### Can I use UV Unfold for models I plan to texture in Substance Painter? Yes. Run UV Unfold in 3D AI Studio, download the model with its new UV layout, and import it into Substance Painter or any other texturing tool. The generated UVs follow standard conventions with properly packed UV islands, so external tools work with them just as well as the built-in texturing tools. --- # Model Actions Source: https://docs.3daistudio.com/utilities/model-actions After you generate a 3D model (or upload one), you can process it further using Model Actions. Open your model from the Dashboard and click the process button to see all available options. ![Model Actions panel showing all available processing options](https://docs.3daistudio.com/assets/images/docs/proc-model-actions.jpg) *The Model Actions panel with all processing options for a 3D model.* ## Available Actions | Action | What it does | Credits | More info | | --- | --- | --- | --- | | **Retopology** | Rebuild mesh with clean topology, reduce poly count, convert formats | 0-15 | [Remesh Tool](https://docs.3daistudio.com/processing/remesh) | | **Change Format** | Convert between GLB, FBX, OBJ, STL, USDZ, 3MF, PLY | 1 | [Format Conversion](https://docs.3daistudio.com/utilities/format-conversion) | | **Export for 3D Printing** | Watertight STL or multi-color 3MF at real-world size | Free to prepare, download on paid plans | [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing) | | **Repair & Fix Mesh** | Fix holes, normals, and non-manifold geometry | 2-4 | [Mesh Repair](https://docs.3daistudio.com/utilities/mesh-repair) | | **Mesh Segmentation** | AI splits model into logical parts (limbs, clothing, etc.) | 10-15 | [Segmentation](https://docs.3daistudio.com/processing/segmentation) | | **Generate new Texture** | AI retexturing with PBR materials from text or image | 20-30 | [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) | | **Texture Painter** | Paint directly on the 3D model in the browser | Free | [Texture Painter](https://docs.3daistudio.com/textures/texture-painter) | | **Smart Low Poly** | AI retopology to a clean, lightweight mesh; optional UV unwrap in the same step | 15-35 | [Smart Low Poly](https://docs.3daistudio.com/processing/smart-low-poly) | | **UV Unfold** | Clean UV layout; Smart UV keeps textures, rig and animation | 5-15 | [UV Unfold](https://docs.3daistudio.com/processing/uv-unfold) | | **Rig & Animate** | Add a skeleton and animations to character models | 15-20 | [Rigging & Animation](https://docs.3daistudio.com/processing/rigging) | | **Remix Model** | Generate AI variations of the same model | 4-8 | Dashboard only | | **Compress File** | Draco compression and texture optimization | Free | [GLB Compression](https://docs.3daistudio.com/utilities/glb-compression) | | **Add Base** | Add a pedestal or platform under your model | Free | [Add Base](https://docs.3daistudio.com/utilities/add-base) | | **Community Creations** | Submit your model to the public community library | Free | Dashboard only | | **Add LOD's** | Generate Level of Detail variants at different poly counts | 1 per LOD | Dashboard only | ## Image Actions When you generate or upload an image, a different set of actions is available: | Action | What it does | Credits | | --- | --- | --- | | **Use in Image to 3D** | Send the image to Image to 3D to generate a 3D model | Free (generation costs apply separately) | | **Open in Image Studio** | Open the image in Image Studio for further editing | Free | | **Edit Image with AI** | AI-powered editing via prompt (change style, objects, background) | 4-8 | | **Upscale Image** | AI 2x upscale with optional face enhancement | 2 | | **Remove Background** | Remove the background from the image | 2 | | **Change Format & Resize** | Convert between JPG, PNG, WebP with quality settings | Free | | **Delight Image** | Remove harsh shadows and normalize lighting (great prep for 3D conversion) | 2 | | **Community Creations** | Submit the image to the public community library | Free | ## How to Access 1. Open your model from the [Dashboard](https://www.3daistudio.com/Dashboard) 2. Click the process/actions button 3. Choose the action you want 4. Configure settings if needed 5. Results are saved to your model page Most actions require a paid plan. Free tools (Export for 3D Printing, Texture Painter, Compress File, Add Base, Community Creations) run client-side or cost zero credits but still need an active subscription for downloads. ## Frequently Asked Questions ### What can I do with a 3D model after generating it? 3D AI Studio gives you over a dozen processing options right from your Dashboard. You can convert formats, reduce poly count with retopology, add a skeleton and animations, generate new textures, segment the mesh into parts, repair geometry issues, compress the file, and more. Just open your model and click the process button to see every available action. ### Do model actions cost credits? It depends on the action. Several tools in 3D AI Studio are free, including Export for 3D Printing, Texture Painter, GLB Compression, and Add Base. Other actions range from 1 credit (format conversion) up to 30-35 credits for more intensive AI processing like texture generation or smart low poly. The credits column in the actions table above shows the exact cost for each. ### Can I process a model I uploaded myself, or only AI-generated ones? You can use model actions on any model in your 3D AI Studio library, whether it was generated with Image to 3D, Text to 3D, or uploaded from your own files. Upload your model through the Dashboard, and all the same processing tools become available. ### Which model action should I use to prepare a model for Unity or Unreal Engine? Start with Format Conversion to export as FBX, which is the preferred format for both Unity and Unreal Engine. If the poly count is too high, run the Retopology action first to bring it down. 3D AI Studio also has dedicated integration guides for [Unity](https://docs.3daistudio.com/integrations/unity) and [Unreal Engine](https://docs.3daistudio.com/integrations/unreal) with engine-specific tips. --- # Format Conversion Source: https://docs.3daistudio.com/utilities/format-conversion Format Conversion turns any 3D model in your 3D AI Studio library into GLB, GLTF, FBX, OBJ, STL, USDZ, 3MF or PLY. Each conversion costs 1 credit, takes a few seconds, keeps the original file and does not change the mesh or textures. | | | | --- | --- | | **Cost** | 1 credit per conversion | | **Time** | A few seconds | | **Input** | Any model in your dashboard | | **Output formats** | GLB/GLTF, FBX, OBJ, STL, USDZ, 3MF, PLY | | **Quality loss** | None, geometry and textures are repacked as they are | ![Format conversion dialog showing supported output formats](https://docs.3daistudio.com/assets/images/docs/proc-format-change.jpg) *Select your target format and convert instantly* ## How do I change a model's format? Open the model in your dashboard, choose **Process → Change Format** and pick FBX. The FBX appears on the model page next to the original GLB a few seconds later. 1. Open the model from your [Dashboard](https://www.3daistudio.com/Dashboard) 2. Click **Process** to open the model actions 3. Select **Change Format** 4. Choose **FBX**, or any other target format 5. Download the new file from the model page The same steps convert to OBJ, STL, USDZ, 3MF, PLY or back to GLB. Converting a file that is not in your 3D AI Studio library? The free [GLB to FBX converter](https://www.3daistudio.com/convert/glb-to-fbx) runs in your browser with no signup, and the file never leaves your computer. The same [3D file converter](https://www.3daistudio.com/convert) handles [GLB to STL](https://www.3daistudio.com/convert/glb-to-stl), [STL to 3MF](https://www.3daistudio.com/convert/stl-to-3mf), [FBX to GLB](https://www.3daistudio.com/convert/fbx-to-glb) and other format pairs. ## Which format should I convert to? Pick the format your destination expects: FBX for Unity and Unreal, GLB or OBJ for Blender, STL for single-color printing, 3MF for color printing and USDZ for AR on Apple devices. | Destination | Format | Why | | --- | --- | --- | | Unity, Unreal Engine | **FBX** | Native import with materials, rigging and animation | | Blender | **GLB** or **OBJ** | GLB keeps textures in one file, OBJ works everywhere | | Web viewers, Godot | **GLB** | Compact single file, the web standard | | 3D printing, one color | **STL** | Accepted by every slicer | | 3D printing, several colors | **3MF** | Stores filament colors for AMS and MMU units ([Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing)) | | iPhone, iPad, Vision Pro AR | **USDZ** | Opens in AR Quick Look | | Point cloud or scan tools | **PLY** | Stores vertex colors | The [Export Formats guide](https://docs.3daistudio.com/export-formats) compares every format in detail, including GLB vs FBX and STL vs 3MF. ## Does converting reduce quality? No. Conversion repacks the existing geometry and textures into the new format without changing them. To lower the polygon count or rebuild the topology, use the [Remesh Tool](https://docs.3daistudio.com/processing/remesh), which can also export in a new format in the same step. ## Can I convert several models at once? Not yet. Conversions run one model at a time from the dashboard. Each one takes a few seconds, so converting a handful of models is still quick. To download many converted files together, use [ZIP Export](https://docs.3daistudio.com/utilities/zip-export). ## Related tools - [Remesh Tool](https://docs.3daistudio.com/processing/remesh): convert the format and reduce the polygon count in one step - [Export Formats](https://docs.3daistudio.com/export-formats): which format to use for each project - [Blender](https://docs.3daistudio.com/integrations/blender), [Unity](https://docs.3daistudio.com/integrations/unity) and [Unreal Engine](https://docs.3daistudio.com/integrations/unreal) import guides - Generate models to convert with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) or [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d) --- # Mesh Repair Source: https://docs.3daistudio.com/utilities/mesh-repair The Repair and Fix Mesh tool automatically detects and corrects common geometry problems in your 3D models. AI-generated meshes sometimes have small defects like holes, flipped normals, or non-manifold edges that cause issues in game engines, slicers, or other software. This tool fixes those problems so your model is ready to use. ![Mesh repair dialog showing repair options](https://docs.3daistudio.com/assets/images/docs/proc-repair.jpg) *Choose between geometry-only repair or full repair with texture re-bake* ## What it fixes - **Non-manifold geometry** - Edges shared by more than two faces, or vertices with complex topology - **Holes and gaps** - Small openings in the mesh surface get filled in - **Flipped normals** - Faces pointing the wrong direction are corrected - **Degenerate faces** - Zero-area triangles and duplicate vertices are cleaned up - **Self-intersections** - Overlapping geometry is resolved where possible ## Repair modes **Geometry only (2 credits)** - Fixes the mesh structure without touching textures. Best when you plan to retexture the model or only need a clean mesh for further processing. **With texture re-bake (4 credits)** - Repairs the geometry and then re-bakes the textures onto the fixed mesh. This ensures the UV mapping still looks correct after vertices are moved or added. Use this mode when you want a ready-to-use result. ## How to use 1. Open a model from your [Dashboard](https://www.3daistudio.com/Dashboard) 2. Click **Process** and select **Repair & Fix Mesh** 3. Choose your repair mode 4. The repaired model replaces the original or is added as a new version Mesh Repair is available as a model action from the [3D AI Studio Dashboard](https://www.3daistudio.com/Dashboard). ## Related Tools - Need to reduce poly count too? Use [Remesh](https://docs.3daistudio.com/processing/remesh) for retopology and optimization - Preparing for 3D printing? [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing) makes the model watertight and exports single or multi-color print files - Need to retexture after repair? Use the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) - See all available model actions in the [Model Actions overview](https://docs.3daistudio.com/utilities/model-actions) ## Frequently Asked Questions ### How do I fix holes in a 3D model? Open your model in 3D AI Studio, click Process, and select Repair & Fix Mesh. The tool automatically detects and fills holes, fixes flipped normals, and resolves non-manifold geometry. Choose the "with texture re-bake" option (4 credits) if you want the textures to stay aligned after the repair, or geometry-only (2 credits) if you plan to retexture later. ### What does non-manifold geometry mean, and why does it matter? Non-manifold geometry means edges or vertices that create impossible shapes, like an edge shared by three or more faces. This causes problems in 3D printing slicers, game engines, and rendering software. 3D AI Studio's Mesh Repair tool fixes non-manifold issues automatically so your model works correctly in downstream applications. ### Can I repair a model for 3D printing? Yes, 3D AI Studio's Mesh Repair is a great first step before printing. It seals holes, fixes normals, and cleans up degenerate faces that would cause slicer errors. For printing, [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing) also seals the model automatically and exports a print-ready 3MF or STL. ### Does mesh repair change how my model looks? In most cases the visual difference is minimal. The tool targets structural defects, not the overall shape. If vertices need to move slightly to fix topology issues, the "with texture re-bake" mode in 3D AI Studio re-maps the textures onto the repaired mesh so everything still looks correct. --- # 3D Viewer Source: https://docs.3daistudio.com/utilities/3d-viewer The 3D Viewer is a free in-browser tool for viewing and inspecting 3D files. Drag and drop a model to view it instantly with realistic lighting and full orbit controls. No account needed, no credits used, and your files never leave your computer. ![3D Viewer showing a model with HDRI environment lighting](https://docs.3daistudio.com/assets/images/docs/proc-viewer.jpg) *A model loaded in the 3D Viewer with HDRI environment lighting* ## Supported formats - GLB / GLTF - FBX - OBJ (with MTL) - STL ## Features **HDRI environment lighting** - Choose from several environment maps to see how your model looks under different lighting conditions. The HDRI provides realistic reflections on metallic and glossy surfaces. **Material inspection** - View individual material channels including albedo, normal map, roughness, and metallic. This helps you verify that textures exported correctly. **Screenshot export** - Capture a high-resolution screenshot of your current view. Useful for sharing previews or creating thumbnails. **Orbit, pan, and zoom** - Left-click to orbit, right-click to pan, scroll to zoom. The viewer centers automatically on your model. ## Privacy All processing happens locally in your browser. Your files are not uploaded to any server. This makes the viewer suitable for confidential or pre-release work. ## Related Tools - Want a spinning video of your model? Use the [Model Spinner](https://docs.3daistudio.com/utilities/model-spinner) to render turntable videos - Need product photos? The [3D Stager](https://docs.3daistudio.com/utilities/3d-stager) places your model in AI-generated scenes - Generate models to view with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) or [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d) - See all [Export Formats](https://docs.3daistudio.com/export-formats) supported by 3D AI Studio ## Frequently Asked Questions ### Can I view a GLB file online without installing software? Yes. 3D AI Studio's free 3D Viewer runs entirely in your browser with no installation, no account, and no file uploads. Just drag and drop your GLB, GLTF, FBX, OBJ, or STL file and you can orbit, pan, and zoom around the model with realistic HDRI lighting. ### Is the 3D Viewer free to use? Completely free, with no credit cost and no account required. 3D AI Studio processes everything locally in your browser, so your files never leave your computer. This makes it safe for confidential or pre-release models. ### Can I inspect textures and materials in the viewer? Yes. The 3D AI Studio viewer lets you switch between material channels including albedo, normal map, roughness, and metallic. This is useful for checking that your texture export is correct before importing into a game engine or renderer. ### How do I take a screenshot of my 3D model? Load your model in the 3D AI Studio 3D Viewer, position the camera where you want it, and click the screenshot button to capture a high-resolution image. You can use this for quick previews, portfolio thumbnails, or sharing with collaborators. --- # Model Spinner Source: https://docs.3daistudio.com/utilities/model-spinner The Model Spinner creates smooth 360-degree turntable videos from your 3D models. Upload a GLB file, customize the rotation, and download a video showing your model from every angle. It is free to use with no credits required. ![Model Spinner interface with a rotating model preview](https://docs.3daistudio.com/assets/images/docs/proc-spinner.jpg) *Set your background, speed, and angle before rendering the video* ## How to use 1. Upload a GLB file or select a model from your Dashboard 2. Set the rotation speed and direction 3. Choose a background color or enable transparent background 4. Click render and download your video ## Options **Background** - Pick a solid color, gradient, or transparent (alpha channel). Transparent backgrounds are useful when you want to composite the video over other content. **Rotation speed** - Control how fast the model spins. Slower speeds work better for detailed models where viewers need time to appreciate the surface. **Camera angle** - Adjust the vertical angle to show more of the top or sides of your model. ## Output The spinner renders an MP4 video file. When you choose a transparent background, the output includes an alpha channel so you can layer it in video editing software. ## Related Tools - Need AI-rendered product photos instead? Use the [3D Stager](https://docs.3daistudio.com/utilities/3d-stager) - Want to view models interactively? Use the [3D Viewer](https://docs.3daistudio.com/utilities/3d-viewer) - Need to compress models before sharing? Try [GLB Compression](https://docs.3daistudio.com/utilities/glb-compression) - Generate models to showcase with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) or [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d) ## Frequently Asked Questions ### How do I create a 360 turntable video of a 3D model? Upload a GLB file to 3D AI Studio's Model Spinner tool, set your preferred rotation speed and camera angle, choose a background color or transparent option, and hit render. The tool outputs a smooth MP4 video showing your model from every angle. ### Can I make a turntable video with a transparent background? Yes. 3D AI Studio's Model Spinner supports alpha channel output. Select the transparent background option before rendering, and the resulting MP4 will have a clear background so you can composite it over any content in your video editing software. ### Is the Model Spinner free? The Model Spinner in 3D AI Studio is free to use with no credit cost. You can adjust settings and render as many videos as you need. A paid plan is only required if you want to download models from the Dashboard. ### What video format does the spinner export? The Model Spinner exports MP4 videos. When you enable transparent background, the MP4 includes an alpha channel. 3D AI Studio renders the video directly in your browser, so you get the file instantly without waiting for server-side processing. --- # GLB Compression Source: https://docs.3daistudio.com/utilities/glb-compression The GLB Compression tool reduces your 3D model file sizes using Draco mesh compression and texture optimization. Smaller files load faster on the web, use less bandwidth, and improve the experience for viewers on mobile devices. ![GLB Compression interface showing file size reduction](https://docs.3daistudio.com/assets/images/docs/proc-compress.jpg) *Upload a GLB file and see the compressed result with size comparison* ## How it works The tool applies two types of optimization: **Draco mesh compression** - Encodes the geometry data (vertices, normals, UVs) into a compact format. Most 3D viewers and game engines support Draco decoding natively. **Texture optimization** - Reduces texture resolution and applies efficient compression while preserving visual quality. This often provides the biggest file size savings. ## Typical results Most models see a 50-80% reduction in file size. The visual difference is minimal for the majority of use cases, especially for web display. ## Rate limits - **Free users** - 2 compressions per week - **Paid plans** - 10 compressions per day ## How to use 1. Open the tool and upload a GLB file 2. The tool compresses the model automatically 3. Compare the before and after file sizes 4. Download the compressed version ## Related Tools - Need to reduce poly count, not just file size? Use the [Remesh Tool](https://docs.3daistudio.com/processing/remesh) - Preparing for web? GLB is the best format for web viewers and AR. See [Export Formats](https://docs.3daistudio.com/export-formats) - Need to convert to other formats? Use [Format Conversion](https://docs.3daistudio.com/utilities/format-conversion) - Deploying to a game engine? Check the [Integrations guides](https://docs.3daistudio.com/integrations) for engine-specific tips ## Frequently Asked Questions ### How much can I compress a GLB file? Most models see a 50-80% reduction in file size when using 3D AI Studio's GLB Compression tool. The exact savings depend on the model's geometry complexity and texture resolution. Visual quality stays nearly identical for web display and most other use cases. ### What is Draco compression? Draco is an open-source compression library by Google that encodes 3D mesh data (vertices, normals, UVs) into a much smaller format. 3D AI Studio applies Draco compression along with texture optimization to minimize your GLB file size. Most modern 3D viewers, game engines, and web frameworks support Draco decoding natively. ### Is there a limit on how many files I can compress? Free users get 2 compressions per week, while paid plans on 3D AI Studio allow 10 compressions per day. Each compression runs automatically once you upload a GLB file, and you can download the optimized version right away. ### Will compressed GLB files work in game engines and web viewers? Yes. Draco-compressed GLB files are supported by Three.js, Babylon.js, Unity, Unreal Engine, and most modern 3D tools. 3D AI Studio's compression produces standard Draco-encoded GLB files that load faster without compatibility issues. --- # Add Base Source: https://docs.3daistudio.com/utilities/add-base The Add Base tool places a pedestal or platform underneath your 3D model. This is useful for display purposes, 3D printing stability, or creating a more polished presentation. The processing runs entirely in your browser. ![Add Base interface showing a model with a circular pedestal](https://docs.3daistudio.com/assets/images/docs/proc-add-base.jpg) *A model with a round base added for display* ## How to use 1. Upload a GLB file or select a model from your Dashboard 2. Choose a base shape (round, square, or hexagonal) 3. Adjust the size and height of the base 4. Preview the result in the built-in viewer 5. Download the model with the base attached ## Options **Shape** - Round, square, or hexagonal base shapes. Round works well for most organic models. Square or hexagonal bases suit architectural or mechanical objects. **Size** - Scale the base relative to your model. The tool auto-fits by default but you can override this. **Height** - Control how tall the pedestal is. Thinner bases work for display, thicker ones add stability for 3D printing. ## Pricing Processing is free and runs client-side in your browser. Downloading the result requires a paid plan. ## Related Tools - Preparing for 3D printing? [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing) maps the colors to your filaments and sets the print size - Need product showcase photos? Use the [3D Stager](https://docs.3daistudio.com/utilities/3d-stager) to place your model in scenes - Want a spinning turntable video? Try the [Model Spinner](https://docs.3daistudio.com/utilities/model-spinner) - Generate models to add bases to with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) or [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d) ## Frequently Asked Questions ### How do I add a pedestal or stand to a 3D model? Open the Add Base tool in 3D AI Studio, upload a GLB file or pick a model from your Dashboard, and choose from round, square, or hexagonal base shapes. Adjust the size and height to fit your model, preview the result, and download the combined file. ### Which base shape works best for 3D printing? For most figurines and organic models, a round or hexagonal base provides even support and looks clean. Square bases work well for architectural or boxy models. In 3D AI Studio, you can also increase the base height for extra printing stability, especially for models with small or uneven contact points. ### Does adding a base cost credits? No. The Add Base tool in 3D AI Studio is completely free and runs in your browser. The base is merged directly into the model geometry, so no server processing is needed. You just need a paid plan to download the final result. ### Can I add a base and then export for 3D printing? Yes, and that is a common workflow. Add your base in 3D AI Studio first, then open [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing) to pick your filament colors and export as 3MF or STL. The pedestal exports as part of the model, so it prints as one solid piece. --- # Export for 3D Printing Source: https://docs.3daistudio.com/utilities/export-for-3d-printing Export for 3D Printing turns any 3D model into a print-ready file in your browser. It seals the model into a watertight solid, lets you print it in one color, with up to 8 filaments or in full color, and downloads a 3MF that opens in Bambu Studio, OrcaSlicer or PrusaSlicer with the colors already painted in. | | | | --- | --- | | **Cost** | Preparing and previewing are free, downloading is included in paid plans | | **Runs** | On your device, the model is never uploaded | | **Opens from** | Any Dashboard model, Change Format, or a GLB/GLTF upload | | **Print setups** | Single color (any printer), Multi-color with 1-8 filaments (AMS, MMU, tool changer), Full color (ProJet, PolyJet, Mimaki, HP) | | **Preparation** | Automatically watertight and checked edge by edge | | **Downloads** | 3MF, STL, OBJ, GLB, plus VRML and PLY for full color | ![Export for 3D Printing showing a fantasy warrior model with print setup, DetailForge, printer, height and file format settings](https://docs.3daistudio.com/assets/images/docs/export-for-3d-printing.webp) *Export for 3D Printing: the print preview on the left, and the print setup, DetailForge, printer, height and file format on the right.* Export for 3D Printing replaced the earlier vertex color tool on October 1, 2026. The old workflow is still described on the [Vertex Color (Legacy)](https://docs.3daistudio.com/utilities/vertex-color) page. ## How do I prepare a 3D model for 3D printing? Open the model, choose **Export for 3D Printing**, pick single or multi-color, set the height and download. The model is made watertight automatically, so it opens cleanly in your slicer without a separate repair step. 1. Open a model from your [Dashboard](https://www.3daistudio.com/Dashboard) and choose **Export for 3D Printing** in its actions. To use a file from elsewhere, upload a GLB or GLTF to the [standalone tool](https://www.3daistudio.com/Tools/VertexColor) 2. Wait a few seconds while the model is sealed into one watertight solid, verified edge by edge and its colors are read 3. Choose **Single color**, **Multi-color** or **Full color**, and switch the preview between **Original**, **Compare** and **Printed** to see how the print will look 4. Set the height in millimeters and choose your printer 5. Optionally turn on **DetailForge** for sharper surface detail 6. Pick a format and download The **Change Format** dialog also links here when you pick a 3D printing format, so you can choose colors and size before downloading. ## How do I 3D print in multiple colors? Choose **Multi-color**, pick how many filaments you have (1 to 8) and download a **3MF**. The model's own colors are turned into clean, printable color regions that still look like the original, and the slicer receives them as filaments 1, 2, 3 and so on. - **Suggested colors:** colorful models start with 4 filaments, what most AMS and MMU units hold, and the tool suggests the number of colors that captures your model best - **Color detail:** **Clean** makes big, calm color areas with the fewest filament swaps, **Balanced** keeps faces and markings readable with few swaps, **Fine** follows the texture closely with more swaps and purge - **Protect small details:** eyes, mouths and markings keep a contrasting filament instead of vanishing, even with just a few colors - **Cleanup:** specks and lines too thin for your nozzle are removed, so fewer color changes are wasted While you adjust, the tool estimates what the colors cost at the printer: | Estimate | What it tells you | | --- | --- | | **Looks like the model** | How closely the printed colors match the original, in percent | | **Filament swaps** | Roughly how many color changes the print needs | | **Purge** | Roughly how much filament the swaps waste, in grams | | **Extra time** | Roughly how much longer the color changes make the print | The estimates assume an AMS or MMU at 0.2 mm layers. A tool changer swaps faster. Multi-color is only available when the model has colors of its own. ## How do I match the colors to my filaments? Save the filaments you own, then let the export use them or match the colors to real filaments in one click. You can still change any color by hand and go back with **Reset colors**. - **Your filaments:** saved in your browser and offered first every time - **Match to filaments:** maps each color to the closest of your saved filaments, or to common filament colors if you have not saved any - **Change a color:** pick any color for a region by hand - **Hover to find:** hovering a color highlights where it is on the model ## How do I fix spots in the wrong color? Use **Touch up** to paint over areas that ended up in the wrong filament. It works on multi-color exports with two or more colors, and your paint is included in the downloaded file. Pick the filament to paint with, adjust the brush size and paint directly on the preview. Undo and redo work as usual, and **Done** returns to the export settings. ## How do I print in a single color? Choose **Single color** and pick your filament color from the quick colors, your saved filaments or a custom color. The color only changes the preview and the 3MF. STL files carry no color. On a colorful model, the tool offers to switch to Multi-color if you have an AMS or MMU. ## How do I export for full color 3D printing? Choose **Full color** to keep every shade of the model instead of reducing it to a few filaments. The file carries the original colors at full texture resolution for full-color printers such as 3D Systems ProJet, Stratasys J-series (PolyJet), Mimaki and HP Jet Fusion color, and for full-color print services. The printer list then shows full-color printers and their software, for example 3D Sprint for ProJet, GrabCAD Print for Stratasys and Mimaki 3D Print prep for Mimaki. Pick one, or keep **Any printer (not listed)** for a print service. | Format | Contains | Best for | | --- | --- | --- | | **OBJ** | Model and texture in a .zip | Almost every full-color printer and print service | | **VRML (.wrl)** | Model and texture in a .zip | 3D Systems ProJet and ZPrinter | | **3MF** | One file with the texture inside | Stratasys GrabCAD Print, Mimaki and HP | | **PLY** | One file with a color per vertex | Smaller files, but less sharp than a texture | Unzip OBJ and VRML downloads and keep the texture next to the model file. Full color is only for printers that print color directly: filament (FDM) printers can't print every shade, so use **Multi-color** for an AMS or MMU instead. ## How do I set the print size? Enter the exact height in millimeters and choose your printer. The tool checks that the model fits the build plate, and the downloaded file arrives at that size in millimeters, standing on the plate. The printer list covers popular models from Bambu Lab, Prusa, Creality, Anycubic, Elegoo and other brands. It also knows each printer's slicer and multi-color unit, for example Bambu Studio with an AMS, so the tips after download match your setup. ## What is DetailForge? DetailForge is an optional step that turns surface detail into real geometry. **Texture detail** engraves painted-on details like outlines, seams and patterns as grooves you can feel, even on a one-color print. **Sharpen** crisps the creases and edges already sculpted into the model. | Setting | What it does | | --- | --- | | **Texture detail** | Engraves the painted lines. Fine (0.35 mm lines) suits resin and large prints, Medium (0.5 mm) suits most models, Bold (0.8 mm) suits small filament prints | | **Depth** | How deep the lines go, from 0.1 to 1.5 mm | | **Raise the lines** | Makes the lines stand out instead of cutting in | | **Sharpen** | Deepens creases and crisps bevels, up to 300% | Hold **compare** to see the model with and without DetailForge before you decide. ## Which file format should I download? Download **3MF** for multi-color printing, because it is the only format that carries the filament colors to your slicer. Use **STL** for single-color printing in any slicer. | Format | Multi-color | Single color | | --- | --- | --- | | **3MF** | Colors painted in, ready for Bambu Studio, OrcaSlicer and PrusaSlicer | Keeps the size and color, opens centered on your printer's plate | | **STL** | Shape only | Works with every slicer | | **OBJ** | Colors per vertex, for 3D apps and full-color print services | For 3D apps like Blender | | **GLB** | With colors, for viewers, AR and the web | For games, AR and web viewers | For full-color printers, the formats are OBJ, VRML, 3MF and PLY, as described [above](https://docs.3daistudio.com/utilities/export-for-3d-printing#how-do-i-export-for-full-color-3d-printing). 3MF files can also be previewed in your dashboard, including the colors of multi-color print files. For a deeper comparison, see [3MF vs STL](https://www.3daistudio.com/convert/compare/3mf-vs-stl). ## Is Export for 3D Printing free? Preparing the model and previewing the print are free for everyone. Downloading the print files is included in the paid plans, so you can check the result before deciding. ## Related tools - [Add Base](https://docs.3daistudio.com/utilities/add-base): add a pedestal or stand before printing - [Mesh Repair](https://docs.3daistudio.com/utilities/mesh-repair): fix a broken mesh with more control - [Smart Low Poly](https://docs.3daistudio.com/processing/smart-low-poly) and [Remesh](https://docs.3daistudio.com/processing/remesh): reduce the polygon count of heavy models - [Export Formats](https://docs.3daistudio.com/export-formats): every format compared - Need a model to print? Generate one with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) or [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d), or follow the step-by-step article on [how to 3D print a picture](https://www.3daistudio.com/blog/how-to-3d-print-a-picture-from-2d-to-3d-in-easy-steps) ## Frequently Asked Questions ### Why are Multi-color and Full color unavailable for my model? Both need a model with colors of its own, such as textures or vertex colors. A model without any colors can only be exported in a single color. To print it in color, texture it first with the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator), then open Export for 3D Printing again. ### Do I need to repair the model before exporting? No. Export for 3D Printing seals every model into a watertight solid and checks it edge by edge before you choose any settings, and it tells you how many problems it fixed. Use [Mesh Repair](https://docs.3daistudio.com/utilities/mesh-repair) only when you want to fix a model for other uses too. --- # ZIP Export Source: https://docs.3daistudio.com/utilities/zip-export ZIP Export bundles any of your generations - images, 3D models, and videos - into a single downloadable archive. Instead of saving files one by one, select everything you need and download it as one ZIP. ## Where you can export **From your Dashboard:** 1. Open your [Dashboard](https://www.3daistudio.com/Dashboard) 2. Select the files you want (Ctrl-click for several, Shift-click for a range - selections work across pages) 3. Click **Export ZIP** in the action bar 4. Pick your options and start the export **From Flow:** - Add the **Export ZIP** node (Logic category) at the end of a workflow. Wire any results into it - single files, whole batches, or a Collect - and run the flow. When everything finishes generating, your ZIP is created automatically. [Learn more about nodes](https://docs.3daistudio.com/flow/nodes) - Or use the **Outputs panel**: click the select button (or Ctrl-click any result) to pick up to 50 outputs and export them directly. ![The Export ZIP node in Flow with connected results, options, and a finished download](https://docs.3daistudio.com/assets/images/docs/flow/flow-export-zip-node.webp) *The Export ZIP node: wire in results, pick your options, and download the finished archive* ## Export options | Option | What it does | | --- | --- | | **Name** | Becomes the ZIP's filename (e.g. "Game assets pack.zip"). Optional. | | **Folders** | Organize the archive: by type (2D / 3D / video), by file format (glb, png, ...), or by date. Default is no folders. | | **Link duration** | How long the download stays available: **1 day** (default) or **30 days**. | | **Email me** | Get an email with the download link the moment your export is ready - handy for larger exports. | ## Tracking your exports ![The My Exports modal listing finished exports with sizes, expiry times, and download buttons](https://docs.3daistudio.com/assets/images/docs/dashboard-my-exports.webp) *My Exports: every export with its size, expiry countdown, and download button* The **My Exports** button in your Dashboard header shows everything in one place: - A live progress bar while an export is being prepared (you can cancel anytime) - A **spinner** on the button while an export runs, and a **green dot** when one is ready that you haven't seen yet - Download links for every export that's still within its link duration You don't need to keep the page open - start an export, keep working, and come back for it (or let the email find you). ## Good to know - Each export can contain up to **50 files** - Each export costs **1 credit**. If an export fails or you cancel it, the credit is refunded automatically - If some selected files are no longer available, the export still completes with everything else - the archive includes a note listing what was skipped - Once a download link expires, the ZIP is gone for good. Your original files stay untouched in your Dashboard - just export them again if you need a new link - 3D models that were exported as OBJ or FBX bundles (mesh + textures) stay bundled inside the archive ## Related Tools - Choosing the right 3D file format? See [Export & Formats](https://docs.3daistudio.com/export-formats) - Convert models before exporting with [Format Conversion](https://docs.3daistudio.com/utilities/format-conversion) - Build automated pipelines that end in a ZIP with [Flow](https://docs.3daistudio.com/flow/overview) ## Frequently Asked Questions ### How do I download all my 3D models at once? Open your Dashboard, select the models you want (selections work across pages), and click Export ZIP. You'll get a single archive with everything inside - up to 50 files per export. ### How long is my ZIP download available? You choose when you create the export: 1 day or 30 days. After that the link expires, but your files stay in your Dashboard and you can export them again anytime. ### Can I organize the ZIP into folders? Yes. Choose "Folders by type" to split 2D, 3D, and video into separate folders, "Folders by format" to group by file type (glb, png, mp4, ...), or "Folders by date" to group by creation day. ### Does exporting cost credits? Each export costs 1 credit, regardless of how many files it contains. If the export fails or you cancel it, the credit is refunded automatically. --- # 3D Stager Source: https://docs.3daistudio.com/utilities/3d-stager The 3D Stager lets you place your 3D models into virtual scenes, adjust lighting and camera angles, and then render photorealistic product photos using AI. Instead of hiring a photographer or building a physical set, you can generate professional product images directly from your 3D models. ![3D Stager showing a model placed in a scene before rendering](https://docs.3daistudio.com/assets/images/docs/proc-stager-input.jpg) *Position your model in the scene, set the lighting, and choose your camera angle* ## How it works 1. Upload a GLB model or select one from your Dashboard 2. Place it in a scene with a surface and background 3. Adjust the lighting, shadows, and camera position 4. Write a prompt describing the desired look (or use the defaults) 5. Choose an AI model and render The AI takes your staged scene as input and generates a photorealistic image that matches the composition, lighting direction, and model placement you set up. ![AI-rendered product photo output from the 3D Stager](https://docs.3daistudio.com/assets/images/docs/proc-stager-output.jpg) *The AI-rendered result showing a photorealistic product photo* ## AI render models Three rendering models are available, each with different strengths: | Model | Credits | Best for | |-------|---------|----------| | Nano Banana 2 | 4 | Fast renders, good general quality | | Gemini Edit | 4 | Strong prompt adherence, detailed textures | | Nano Banana Pro | 6 | Highest quality, best lighting and reflections | ## Scene controls **Surface** - Choose a flat surface material (wood, marble, concrete, fabric, or custom color). **Lighting** - Adjust intensity, direction, and color temperature. The AI uses your lighting setup as a guide for the final render. **Camera** - Orbit freely or pick preset angles (front, three-quarter, top-down). The composition you see is what the AI receives. **Prompt** - Describe the environment or mood you want. For example: "product on a marble countertop, soft morning light, minimal background." The AI blends your scene setup with the text prompt. ## Tips for best results - Position the model so it fills most of the frame - Keep lighting simple and directional for clean shadows - Use short, descriptive prompts rather than long detailed ones - Try multiple AI models to see which suits your product style ## Related Tools - Want turntable videos instead? Use the [Model Spinner](https://docs.3daistudio.com/utilities/model-spinner) - Need to view models interactively? Try the [3D Viewer](https://docs.3daistudio.com/utilities/3d-viewer) - Want to add textures before staging? Use the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) - Generate models to stage with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) or [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d) ## Frequently Asked Questions ### Can I create product photos from a 3D model without a photographer? Yes. 3D AI Studio's 3D Stager lets you place your model in a virtual scene, set up lighting and camera angles, and render photorealistic product photos using AI. You get professional-looking images in seconds without any physical setup, studio, or camera equipment. ### How much does it cost to render a product image? Each AI render in 3D AI Studio's 3D Stager costs 4-6 credits depending on which rendering model you choose. Nano Banana 2 and Gemini Edit cost 4 credits each, while Nano Banana Pro costs 6 credits and produces the highest quality output with better lighting and reflections. ### What kind of scenes and backgrounds can I create? You can choose from surface materials like wood, marble, concrete, and fabric, then write a text prompt to describe the environment you want. For example, "product on a marble countertop, soft morning light, minimal background." 3D AI Studio's AI blends your scene setup with the prompt to generate the final image. ### Which AI render model should I pick? Start with Nano Banana 2 (4 credits) for quick tests and general use. Try Gemini Edit (4 credits) when you need strong prompt adherence and detailed textures. Use Nano Banana Pro (6 credits) from 3D AI Studio for your final hero shots where lighting quality and reflections matter most. --- # Model Storage & Expiry Source: https://docs.3daistudio.com/utilities/model-storage Every 3D model you generate is stored in your library for a set period, so your Dashboard stays fast and we don't keep files nobody needs anymore. You can always see when a file's storage period ends, and you can extend it. ## How long models are kept | Plan | Storage period | | --- | --- | | Free plan | **6 months** | | Paid plans (Basic, Studio, Pro) and credit packs | **12 months** | | After **Extend storage** | **3 years** | The period is counted from the day the model was created, and the plan that applies is the one you were on at that moment. Models made while you're on a paid plan keep their 12 months even if you later downgrade or cancel - storage is never shortened. Team files follow the team's plan. ## Seeing a file's storage date Open any 3D model in your Dashboard. The **Details** panel shows a **Stored** row with the exact date the file is kept until, for example *Until Mar 6, 2027*. During the last month before that date, the file's card in your Dashboard shows a small clock with the days left, so it's easy to spot files that are about to expire. ## Extending storage Any 3D model that is still available can be extended once: 1. Open the model in your Dashboard 2. In **Details → Stored**, click **Extend storage · 1 credit** 3. Confirm - the row updates to the new date right away Extended files are kept for **3 years from the day they were created**. The confirmation shows the exact resulting date before you pay. Good to know: - Extending costs **1 credit** and can be done **once per file** - Files up to **150 MB** can be extended. For larger files, run them through [GLB Compression](https://docs.3daistudio.com/utilities/glb-compression) first and extend the compressed copy - Each file is extended on its own: converted copies, rigged versions and other derived files have their own storage dates ## What happens when a file expires When the storage period ends, the model file itself is removed and can no longer be viewed, downloaded or processed. Everything around it stays in your library: - The thumbnail and the card in your Dashboard - The input images, prompt and generation settings - Tags, notes and project membership Opening an expired file shows when it was stored until and offers **Generate again** with the same inputs and settings when those were recorded. The expired file can't be restored. ## Images, videos and other files Storage periods apply to 3D model files. Images and videos generated in Image Studio and Flow are not affected. Questions about storage? Contact [support@3daistudio.com](mailto:support@3daistudio.com). --- # Export & Formats Source: https://docs.3daistudio.com/export-formats 3D AI Studio exports models as GLB, FBX, OBJ, STL, USDZ, BLEND, 3MF, Collada and PLY. Use **GLB** by default, **FBX** for Unity and Unreal, **STL** or **3MF** for 3D printing, and **USDZ** for AR on Apple devices. Generate a model with [Image to 3D](https://www.3daistudio.com/ImageTo3D) or [Text to 3D](https://www.3daistudio.com/TextTo3D) first, then export it from your [Dashboard](https://www.3daistudio.com/Dashboard). ## Which format should I download? Pick the format your destination expects. When in doubt, use GLB: it works almost everywhere and keeps geometry, textures and materials in one file. For a general introduction to every format, see the [3D file formats guide](https://www.3daistudio.com/convert/3d-file-formats). | Format | Best for | Textures | Animation | | --- | --- | --- | --- | | **GLB** | Web, AR, Godot, sharing, general use | Embedded | Yes | | **FBX** | Unity, Unreal Engine, rigged characters | Embedded | Yes | | **OBJ** | Blender, ZBrush, maximum compatibility | Separate files (.mtl + images) | No | | **STL** | Single-color 3D printing | No | No | | **3MF** | Multi-color 3D printing | Filament colors | No | | **USDZ** | AR on iPhone, iPad and Vision Pro | Embedded | Yes | | **BLEND** | Native Blender projects | Embedded | Yes | | **Collada (.dae)** | Legacy pipelines | Yes | Yes | | **PLY** | Point clouds and scan data | Vertex colors | No | ## Should I download GLB or FBX? Download GLB for the web, AR, Godot and sharing, and FBX for Unity, Unreal Engine and animation software. Both keep textures and animation, so the choice depends on where the model is going. The [GLB vs FBX comparison](https://www.3daistudio.com/convert/compare/glb-vs-fbx) covers the formats in depth. | | GLB | FBX | | --- | --- | --- | | File structure | One binary file with everything embedded | One file, textures and materials included | | Best for | Web viewers, AR, Godot, sharing | Unity, Unreal Engine, Maya, 3ds Max, Cinema 4D | | Rigging and animation | Supported | Supported, the standard for game pipelines | | Unity | Needs a free package such as GLTFast | Native import, no setup | | Web (Three.js, model-viewer) | Native | Not supported | **Rule of thumb:** use FBX if the model goes into a game engine or animation tool, and GLB for everything else. ## Should I download STL or 3MF for 3D printing? Download STL for single-color prints and 3MF for multi-color prints. STL stores geometry only and works in every slicer, while 3MF also carries the filament colors so printers with filament changers know which color goes where. The [3MF vs STL comparison](https://www.3daistudio.com/convert/compare/3mf-vs-stl) explains the formats in depth. | | STL | 3MF | | --- | --- | --- | | Color | None | Filament colors painted in | | Slicer support | Every slicer (Cura, PrusaSlicer, Bambu Studio) | Modern slicers | | Best for | FDM and resin printing in one color, CNC | Bambu Lab AMS, Prusa MMU and other multi-color printers | | How to get it | Direct download or [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing) | [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing) | ## How do I export a model in a different format? Download GLB directly after generation, or convert to any other format with **Change Format** for 1 credit. If you also need a lower polygon count, use the Remesh tool, which converts and optimizes in one step. For files from other sources, the free in-browser [3D file converter](https://www.3daistudio.com/convert) handles pairs like [GLB to FBX](https://www.3daistudio.com/convert/glb-to-fbx) and [STL to 3MF](https://www.3daistudio.com/convert/stl-to-3mf). | Method | Formats | Cost | Use when | | --- | --- | --- | --- | | **Direct download** in the 3D viewer | GLB (default), STL, USDZ | Free | You only need the file as it is | | **[Change Format](https://docs.3daistudio.com/utilities/format-conversion)** (Model Actions) | GLB, GLTF, FBX, OBJ, STL, USDZ, 3MF, PLY | 1 credit | You need a different format, same mesh | | **[Remesh Tool](https://docs.3daistudio.com/processing/remesh)** | Prism: GLB, FBX, OBJ, STL, USDZ, 3MF. Pro (Meshy): GLB, FBX, OBJ, USDZ, BLEND, STL | Depends on provider | You also want fewer polygons, quads or new topology | To convert with Change Format: 1. Open the model from your Dashboard 2. Click **Process** to open **Model Actions** 3. Select **Change Format** 4. Pick the target format and download the new file from the model page STL and USDZ downloads from the viewer are converted in your browser. Change Format runs on the server and keeps the original file. ## Which format works with my software? | I want to... | Format | Tool | | --- | --- | --- | | Import into Unity or Unreal | FBX | [Change Format](https://docs.3daistudio.com/utilities/format-conversion) or [Remesh](https://docs.3daistudio.com/processing/remesh) | | Edit in Blender | GLB, FBX, OBJ or BLEND | [Change Format](https://docs.3daistudio.com/utilities/format-conversion) | | Use in Godot | GLB | Direct download | | Sculpt in ZBrush | OBJ | [Change Format](https://docs.3daistudio.com/utilities/format-conversion) | | Show on a website or in AR | GLB | Direct download | | View in AR on iPhone or iPad | USDZ | Direct download | | 3D print in one color | STL | Direct download or [Remesh](https://docs.3daistudio.com/processing/remesh) | | 3D print in several colors | 3MF | [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing) | | Build a game asset pipeline | FBX | [Remesh](https://docs.3daistudio.com/processing/remesh) with Prism (quad topology) | | Share a model quickly | GLB | Direct download | Engine-specific steps are in the [Blender](https://docs.3daistudio.com/integrations/blender), [Unity](https://docs.3daistudio.com/integrations/unity), [Unreal Engine](https://docs.3daistudio.com/integrations/unreal) and [Godot](https://docs.3daistudio.com/integrations/godot) guides. ## Format details ### GLB GLB packs geometry, textures and materials into a single binary file. It is small, fast and supported by every modern 3D tool, which makes it the recommended default. **Use for:** web viewers (Three.js, model-viewer), AR and VR, Godot, sharing and general storage. ### FBX FBX is the standard for game engines and professional 3D tools and preserves materials, rigging and animation data. **Use for:** Unity, Unreal Engine, Maya, 3ds Max, Cinema 4D and rigged or animated characters. **Tip:** when exporting FBX from the Remesh tool with Prism, you can choose presets for Blender, 3ds Max or Mixamo. ### OBJ OBJ is the simplest and most widely supported format. It stores geometry in the .obj file and references textures through a separate .mtl file. **Use for:** Blender, ZBrush, Maya and older software where compatibility matters most. **Note:** keep the .obj, .mtl and texture files in the same folder, or use **OBJ ZIP** from the download menu to get everything bundled. ### STL STL stores geometry only, without textures or colors, and is the standard format for 3D printing. **Use for:** FDM printing (Cura, PrusaSlicer, Bambu Studio), resin printing and CNC. **Tip:** check that the model is watertight before printing. The [Remesh tool](https://docs.3daistudio.com/processing/remesh) can fix geometry issues and reduce the polygon count. ### 3MF 3MF is the modern 3D printing format with color support. It carries the filament colors, so multi-color printers such as Bambu Lab AMS and Prusa MMU know which filament to use for each part. **Use for:** multi-color 3D printing and manufacturing. **Tip:** use [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing) to map the model's colors to your filaments, then download the 3MF. ### USDZ USDZ is Apple's AR format. It opens natively on iPhone, iPad and Vision Pro in Safari, Messages and any ARKit app. **Use for:** AR Quick Look on iOS and macOS, visionOS apps and the Apple ecosystem. ### BLEND BLEND is Blender's native format and keeps the complete scene, including materials and modifiers. **Use for:** projects where Blender is your main tool. ## Frequently Asked Questions ### What is the best format for importing into Unity? FBX. Unity reads FBX materials, textures and rigging natively with no extra setup. GLB also works after installing a free package such as GLTFast. ### What is the best format for Unreal Engine? FBX. Drag the FBX into the Content Browser and Unreal imports the mesh, materials, PBR textures, skeleton and animations. For high-poly static meshes in UE5, enable Nanite after import. ### Which format keeps textures in a single file? GLB, FBX, USDZ and BLEND embed textures in the file. GLB is the most portable, so nothing goes missing when you share it. OBJ keeps textures as separate files next to a .mtl file. ### When should I export USDZ? Export USDZ when people should view the model in AR on an iPhone, iPad or Vision Pro, where it opens directly in Safari and Messages. It is a direct download in the 3D viewer. ### What format should I use for websites? GLB. It is the native format of web viewers such as Three.js, model-viewer and Sketchfab and is the default download in 3D AI Studio. Use USDZ in addition if you want AR on Apple devices. --- # Vertex Color Export (Legacy) Source: https://docs.3daistudio.com/utilities/vertex-color > [!WARNING] > This tool has been replaced. Since October 1, 2026, 3D models are prepared for printing with [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing), which adds a live preview, automatic watertight repair, real-world sizing and support for 1-8 filaments. This page documents the previous version for reference. The original 3D Print Export tool turned a textured 3D model into a multi-color print file by converting its texture colors into vertex colors. It mapped those colors to a filament palette and exported 3MF, OBJ or STL for slicers such as Bambu Studio and PrusaSlicer. | | | | --- | --- | | **Status** | Replaced by [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing) on October 1, 2026 | | **Approach** | Texture colors converted to per-vertex colors, then mapped to a filament palette | | **Modes** | Mesh Only and Palette | | **Exports** | 3MF, 3MF Textured, OBJ with vertex colors, STL | | **Extras** | Six pedestals, color-aware mesh simplification | ![Legacy 3D Print Export tool showing vertex color conversion with a filament palette](https://docs.3daistudio.com/assets/images/docs/proc-vertex-color.jpg) *The previous 3D Print Export interface: 3D preview on the left, color palette, filament mapping and export options on the right.* ## Where did each feature go? Every task from the old tool still exists, either in Export for 3D Printing or in a dedicated tool. | Old feature | Now | | --- | --- | | Palette mode (colors mapped to filaments) | **Multi-color** in [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing), with 1-8 filaments, saved filaments and one-click matching | | Mesh Only mode | **Single color** in Export for 3D Printing | | Similar-color warnings | A suggested number of colors and the **Color detail** setting (Clean, Balanced, Fine) | | Six pedestals | [Add Base](https://docs.3daistudio.com/utilities/add-base) | | Color-aware simplification | [Smart Low Poly](https://docs.3daistudio.com/processing/smart-low-poly) or [Remesh](https://docs.3daistudio.com/processing/remesh) before exporting | | 3MF, OBJ, STL export | 3MF, STL, OBJ and GLB in Export for 3D Printing | | 3MF Textured (full-color printing) | **Full color** in Export for 3D Printing, as OBJ, VRML, 3MF or PLY for ProJet, PolyJet, Mimaki and HP | ## What are vertex colors? Vertex colors store one color per corner (vertex) of the mesh instead of in a separate texture image. AI-generated models keep their colors in textures, which most slicers cannot read, so converting textures to vertex colors was how the old tool made models printable in color. Export for 3D Printing still reads textures and vertex colors, but it turns them into clean filament regions and paints them into the 3MF directly, so you no longer manage vertex colors yourself. ## How did the old tool work? The model's textures were converted to vertex colors and its main colors were extracted automatically. In **Palette mode**, every color was mapped to the nearest filament in your palette, and the result was exported as 3MF for the slicer. 1. Open a model from the Dashboard, or upload a GLB 2. Textures were converted to vertex colors and the main colors extracted 3. **Palette mode** mapped colors to your filaments 4. You adjusted each palette slot to match your spools 5. Optionally, you added a pedestal and simplified the mesh 6. You exported 3MF, OBJ or STL ### Mesh Only vs Palette | | Mesh Only | Palette | | --- | --- | --- | | Colors | Original vertex colors kept | Every color mapped to the nearest filament | | Color matching | None | Perceptual (CIE LAB Delta-E) | | Best for | Single-color prints, slicer-side coloring | Bambu Lab AMS, Prusa MMU and similar systems | Colors closer than Delta-E 28 were flagged as too similar to tell apart on a print. ### Export formats | Format | Contained | Used for | | --- | --- | --- | | **3MF** | Mesh and colors in one file | Bambu Lab AMS, Prusa MMU, modern slicers | | **3MF Textured** | Mesh and original texture images | Full-color printing | | **OBJ with vertex colors** | Mesh and per-vertex RGB colors | Slicer import, editing in Blender | | **STL** | Mesh only, no color | Single-color prints | ### Pedestals and simplification The old tool offered six pedestals (Classic Square, Roman Column, Stone Cylinder, Wood Roots, Marble Disc and Cracked Stone), merged into the model so it printed as one piece. Its color-aware simplification reduced the triangle count by a target percentage while keeping color boundaries sharp. ## Frequently Asked Questions ### Can I still use the old vertex color tool? No. Both the Dashboard action and the standalone tool page now open [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing), which covers the same workflow with a live preview, automatic watertight repair and up to 8 filaments. ### Does multi-color export still work with Bambu Lab AMS? Yes. Choose Multi-color in Export for 3D Printing, set your filaments and download the 3MF. It opens in Bambu Studio with filaments 1 to N already painted on, ready for the AMS. --- # Image Studio Overview Source: https://docs.3daistudio.com/image-studio/overview Image Studio is your all-in-one workspace for creating, editing, and converting images with AI. It includes over 100 tools powered by 15+ AI models, and everything runs in your browser. You can generate images from text, edit photos with natural language, convert images into different art styles, and much more. Whether you are creating concept art, preparing images for 3D conversion, or just experimenting with different styles, Image Studio is the place to do it. ![Image Studio interface overview](https://docs.3daistudio.com/assets/images/docs/is-overview.jpg) *The Image Studio landing page with all tool categories visible.* ## Categories Image Studio organizes its tools into nine categories: | Category | What it does | | --- | --- | | **[Generate](https://docs.3daistudio.com/image-studio/generate)** | Create images from text prompts using 15+ AI models. | | **[Edit](https://docs.3daistudio.com/image-studio/edit)** | Modify existing images with AI. Describe what to change in plain text. | | **[Convert](https://docs.3daistudio.com/image-studio/convert)** | Transform images into different art styles like anime, oil painting, pixel art, and more. | | **Image Reference** | Use reference images to guide generation. Keep styles, colors, and compositions consistent across images. | | **Style Reference** | Apply the visual style of one image to another. Great for maintaining a consistent look across a project. | | **Sketch** | Turn rough sketches into polished images, or draw directly in the built-in sketch painter. | | **Utilities** | Practical tools like background removal, upscaling, and image-to-prompt. | | **Video** | Turn static images into short AI-generated videos. | | **Train** | Train custom AI models on your own images for personalized generation. | ## Keyboard Shortcuts These shortcuts work anywhere inside Image Studio: | Shortcut | Action | | --- | --- | | `Ctrl+Enter` / `Cmd+Enter` | Generate image | | `Ctrl+L` | Open Prompt Library | | `Ctrl+S` | Save current prompt to library | | `Ctrl+P` | Open AI Prompt Helper | | `Ctrl+V` | Paste image from clipboard | | `B` | Toggle sidebar | ## Prompt Library The Prompt Library lets you save, organize, and reuse your prompts across all Image Studio tools. - **Save any prompt** you write and come back to it later. - **Organize with collections and tags** so you can find prompts quickly as your library grows. - **Use templates with variables** like `{{color}}` or `{{style}}` to create reusable prompt structures. Swap out the variable each time you use the template. - Prompts are **shared across all tools**, so a prompt you saved in Generate works in Edit, Convert, or anywhere else. Open the Prompt Library anytime with `Ctrl+L`. ## AI Prompt Helper Not sure what to write? The AI Prompt Helper writes and improves prompts for you. Open it with `Ctrl+P`, describe what you want in a few words, and the AI will expand it into a detailed, effective prompt tailored to the model you are using. You can also paste in an existing prompt and ask the AI to improve it, make it more detailed, or adjust the style. ## Next Steps - [Generate images from text](https://docs.3daistudio.com/image-studio/generate) - [Edit images with AI](https://docs.3daistudio.com/image-studio/edit) - [Convert images to different styles](https://docs.3daistudio.com/image-studio/convert) - [Character Sheet Generator](https://docs.3daistudio.com/image-studio/character-sheet) - [AI Video Generation](https://docs.3daistudio.com/image-studio/video) - [Pose Transfer](https://docs.3daistudio.com/image-studio/pose-transfer) ### From Image to 3D Model Image Studio works great as a first step before 3D conversion. Generate or edit a clean reference image, then send it to [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) to create a full 3D model. See the [Recommended Workflow](https://docs.3daistudio.com/3d-generation/recommended-workflow) for step-by-step examples. ## Frequently Asked Questions ### What is Image Studio and what can it do? Image Studio is the AI image workspace inside 3D AI Studio. It includes over 100 tools for generating, editing, converting, upscaling, and animating images, all powered by 15+ AI models from Google, OpenAI, Black Forest Labs, and ByteDance. Everything runs in your browser with no software to install. ### How is Image Studio different from Midjourney or DALL-E? Unlike Midjourney (Discord-based) or DALL-E (single model), 3D AI Studio's Image Studio gives you 15+ models in one interface so you can pick the best model for each task. It also includes editing, style conversion, video generation, and direct 3D model conversion, making it a complete pipeline rather than just a generator. ### Is Image Studio free to use? 3D AI Studio gives you free credits on signup that work across all Image Studio tools. Free users can access most generation models, conversion styles, and utilities. Paid plans add more credits, higher resolutions, video generation, and download access for tools like Character Sheet and Pose Transfer. ### Can I use Image Studio images for 3D model creation? Yes, that is one of the primary workflows. You can generate or edit an image in Image Studio, then send it directly to Image to 3D with one click. 3D AI Studio's Image Studio includes utilities like background removal and lighting restoration specifically designed to prepare images for clean 3D conversion. ### Does Image Studio support keyboard shortcuts? Yes. 3D AI Studio's Image Studio has keyboard shortcuts for all common actions. Press Ctrl+Enter to generate, Ctrl+L to open the Prompt Library, Ctrl+S to save a prompt, Ctrl+P for the AI Prompt Helper, and Ctrl+V to paste an image from your clipboard. Press B to toggle the sidebar. ### What AI models are available in Image Studio? Image Studio in 3D AI Studio includes models from Google (ImageGen 4, Nano Banana series), Black Forest Labs (Flux 2, Flux 2 Pro), OpenAI (GPT Image 2), and ByteDance (SeedDream). Each model has different strengths, from photorealism to text rendering to speed, and costs between 2-20 credits per generation. --- # Generate Source: https://docs.3daistudio.com/image-studio/generate Generate lets you create images from text prompts. Describe what you want, pick a model, and get your image in seconds. Image Studio includes 15+ generation models from Google, Black Forest Labs, OpenAI, and ByteDance, each with different strengths, speeds, and credit costs. ![Generated orc character](https://docs.3daistudio.com/assets/images/docs/is-generate-orc.jpg) *A generated stylized orc character with detailed fantasy background.* ![Generated futuristic motorcycle](https://docs.3daistudio.com/assets/images/docs/is-generate-bike.jpg) *A generated futuristic motorcycle with sleek, sci-fi design.* ![Generated stylized battle axe](https://docs.3daistudio.com/assets/images/docs/is-generate-axe.jpg) *A generated stylized battle axe with fantasy detailing.* ![Generated LEGO orc figurine](https://docs.3daistudio.com/assets/images/docs/is-generate-lego.jpg) *A generated LEGO-style orc minifigure.* ## Available Models | Model | Provider | Credits | Best for | | --- | --- | --- | --- | | ImageGen 4 | Google | 4 | Detailed, realistic, great lighting | | ImageGen 4 Fast | Google | 2 | Quick iterations | | ImageGen 3 | Google | 4 | Advanced detail | | ImageGen 3 Fast | Google | 2 | Quick drafts | | Nano Banana 2 (Gemini 3.1 Flash) | Google | 4-12 | Flexible resolution (0.5K-4K), grounding | | Nano Banana Pro (Gemini 3 Pro) | Google | 8-14 | Highest quality Gemini | | Nano Banana (Gemini 2.5 Flash) | Google | 4-8 | Fast and affordable | | Flux 2 | Black Forest Labs | 4-6 | Natural language prompts, photorealism | | Flux 2 Flex | Black Forest Labs | 6-10 | Guidance and step control | | Flux 2 Pro | Black Forest Labs | 8 | Commercial-grade | | Flux | Black Forest Labs | 2 | Quick text-to-image | | Flux Pro | Black Forest Labs | 4 | Professional quality | | GPT-Image-1 | OpenAI | 2-20 | Variable quality, text rendering | | GPT Image 2 | OpenAI | 3 | Next-gen, flexible | | SeedDream v5 Lite | ByteDance | 5 | Multi-reference blending | Credit costs vary by resolution and settings. The ranges shown reflect minimum to maximum depending on your chosen output size and options. ## Quick Recommendations Not sure which model to pick? Here are some starting points: - **Best overall quality:** ImageGen 4. Handles most subjects well with great lighting and detail. - **Best for photorealism:** Flux 2. Natural-looking photos with great skin tones and textures. - **Cheapest option:** Flux at just 2 credits. Good for quick ideas and rough concepts. - **Best text rendering:** GPT-Image-1 or GPT Image 2. These handle text in images (signs, labels, logos) better than other models. - **Best for 3D conversion:** ImageGen 4. Add "clean white background, single object, centered" to your prompt for images that convert well to 3D models. ## Prompting Tips Different models respond to prompts differently. Here are the key things to know: **Flux models** prefer natural language. Write prompts like you are describing a scene to someone: "A red sports car parked on a rainy street at night, reflections on wet asphalt, cinematic lighting." Avoid tag-style prompts with commas separating keywords. **ImageGen and Nano Banana models** work well with both natural language and more structured, keyword-style prompts. They handle detailed descriptions and specific art direction effectively. **GPT-Image models** excel at following complex instructions. You can be very specific about composition, layout, and text placement. ### General tips - **Aspect ratio matters.** Pick the right shape for your subject. Portraits work better in portrait orientation, landscapes in landscape. - **Resolution affects cost.** Higher resolutions cost more credits. Start with a lower resolution to test your prompt, then generate the final version at full quality. - **Be specific about what you want.** "A sword" gives you a generic sword. "A curved elven longsword with a silver blade and green leather-wrapped grip, fantasy style" gives you something much more interesting. - **Mention the style early.** Put style keywords like "3D render", "oil painting", "photorealistic", or "anime" near the start of your prompt. For more detailed prompting advice, see [Prompt Tips](https://docs.3daistudio.com/3d-generation/prompt-tips). Ready to generate? [Try the AI Image Generator](https://www.3daistudio.com/ImageGeneration/generate) -- free credits on signup. ## Frequently Asked Questions ### Is there a free AI image generator I can use without signing up for a subscription? Yes. 3D AI Studio gives you free credits when you sign up, so you can generate images immediately without a subscription. Free credits work with all 15+ models including ImageGen 4, Flux, and GPT-Image. You only need a paid plan if you want more credits or access to premium video features. ### Which AI image generator is best in 2026? It depends on what you need. For photorealism, Flux 2 is hard to beat. For all-around versatility and lighting quality, ImageGen 4 is the top pick in 3D AI Studio. GPT Image 2 handles text in images and complex instructions better than any other model currently available. ### How does ImageGen 4 compare to Flux 2 for image generation? ImageGen 4 excels at detailed scenes with realistic lighting and works well with both keyword-style and natural language prompts. Flux 2 produces more natural-looking photos with better skin tones and textures but prefers full-sentence prompts. In 3D AI Studio, ImageGen 4 costs 4 credits and Flux 2 costs 4-6 credits depending on resolution. ### How much does AI image generation cost per image? In 3D AI Studio, costs range from 2 credits (Flux basic) to 20 credits (GPT-Image-1 at max settings). Most popular models like ImageGen 4 and Flux 2 cost 4-6 credits per image. Higher resolutions cost more, so testing prompts at lower resolution first saves credits. ### What is the best AI model for generating images to convert into 3D models? ImageGen 4 produces the cleanest images for 3D conversion. Add "clean white background, single object, centered" to your prompt for best results. In 3D AI Studio, you can send generated images directly to Image to 3D with one click, making the workflow seamless. ### What aspect ratio should I use for AI-generated images? Match the aspect ratio to your subject. Portraits and characters work best in portrait orientation (2:3 or 3:4), landscapes in landscape (16:9 or 3:2), and product shots in square (1:1). 3D AI Studio lets you pick from standard ratios or set custom dimensions for each model. ### Can AI image generators render readable text inside images? GPT-Image-1 and GPT Image 2 are the best models for text rendering in 3D AI Studio. They can generate signs, labels, logos, and UI mockups with legible text. Other models like Flux and ImageGen sometimes produce garbled or missing text, so switch to a GPT model when your image needs readable words. ### What is the difference between Nano Banana, Nano Banana 2, and Nano Banana Pro? Nano Banana (Gemini 2.5 Flash) is the fastest and cheapest at 4-8 credits. Nano Banana 2 (Gemini 3.1 Flash) supports flexible resolution from 0.5K to 4K and offers grounding features at 4-12 credits. Nano Banana Pro (Gemini 3 Pro) delivers the highest quality output at 8-14 credits. All three are available in 3D AI Studio and share the same natural language prompting style. ### How long does AI image generation take? Most models in 3D AI Studio generate an image in 3-10 seconds. Faster models like ImageGen 4 Fast and Flux typically finish in under 5 seconds. Higher-quality options like Nano Banana Pro and Flux 2 Pro may take 8-15 seconds depending on resolution. Video generation is slower, usually 30 seconds to a few minutes. ### Can I use AI-generated images commercially? Commercial usage rights depend on the model. Flux 2 Pro and GPT Image 2 are designed for commercial use. 3D AI Studio does not restrict how you use your generated images, but you should check each model provider's terms if you plan to sell or publish the output commercially. --- # Edit Source: https://docs.3daistudio.com/image-studio/edit Edit lets you modify existing images using AI. Upload any image, describe the changes you want in plain text, and the AI applies them. You can change styles, swap objects, fix lighting, remove elements, or transform a rough sketch into a polished render. ![Original image before editing](https://docs.3daistudio.com/assets/images/docs/is-edit-input.jpg) *Original photo uploaded for editing.* ![Edited image - turned into stylized character](https://docs.3daistudio.com/assets/images/docs/is-edit-output.jpg) *The edited result with the person transformed into a stylized character.* ![Original sketch uploaded for AI editing in Image Studio](https://docs.3daistudio.com/assets/images/docs/is-edit-sketch-input.jpg) *A sketch uploaded to the Edit tool.* ![Sketch rendered into stylized 3D character](https://docs.3daistudio.com/assets/images/docs/is-edit-sketch-output.jpg) *The sketch rendered into a polished stylized 3D character.* ## Available Models | Model | Credits | Best for | | --- | --- | --- | | GPT Image 2 Edit | 3 | Precise edits, text rendering, photorealism | | Nano Banana 2 Edit | 4-12 | Flexible resolution, up to 14 refs | | Nano Banana Pro Edit | 8-14 | Highest quality edits | | Gemini Edit | 4-8 | Multi-image edit, quick iterations | | Flux 2 Edit | 4 | Multi-image Flux editing | | Flux 2 Flex Edit | 6-10 | Precision with guidance control | | Flux 2 Pro Edit | 8 | Commercial-grade editing | | SeedDream v5 Lite Edit | 5 | Multi-reference blending, up to 14 images | | Flux Kontext | 3-6 | Combine/edit up to 4 images | ## Key Principles Getting good results from AI editing comes down to a few core ideas: - **Say what to change AND what to keep.** If you only describe the change, the AI might alter things you wanted to preserve. For example: "Change the hat to red, keep everything else the same." - **One change at a time.** Making multiple changes in a single edit can lead to unpredictable results. Do one edit, check the result, then make the next change. - **Match the lighting.** If you are adding or swapping objects, mention the lighting direction and intensity so new elements blend naturally with the existing scene. ## Common Edit Types Here are some of the most popular ways people use the Edit tools: ### Style transfer Upload a photo and ask the AI to "turn this into Pixar 3D style" or "make this look like a watercolor painting." The AI keeps the composition and subject but changes the entire visual style. ### Background changes Describe a new background: "Place this character in a forest at sunset" or simply "remove the background and make it transparent." ### Object swaps Replace specific elements: "Change the sword to a battle axe" or "replace the wooden table with a glass one." ### Relighting Adjust the mood: "Add dramatic side lighting" or "change to warm golden hour lighting from the left." ### Sketch to render Upload a rough sketch and describe what you want: "Turn this sketch into a detailed 3D render of a robot, metallic materials, studio lighting." This works especially well with GPT Image 2 Edit and Nano Banana 2 Edit. ## Tips for 3D Conversion If you plan to convert your edited image to a 3D model using [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d), a few edits can make a big difference: - **Clean up the background.** Remove busy or distracting backgrounds. A clean white or neutral background gives the 3D model generator a much easier job. - **Center the subject.** Make sure the object fills the frame and is centered. Crop out unnecessary empty space. - **Fix any artifacts.** If the original image has AI artifacts, glitchy areas, or extra limbs, clean those up before sending to 3D conversion. It is much cheaper to fix in 2D than to regenerate the 3D model. - **Use consistent lighting.** Flat, even lighting from the front works best for 3D conversion. Avoid dramatic shadows that might confuse the depth estimation. For the full recommended workflow from image to 3D model, see [Recommended Workflow](https://docs.3daistudio.com/3d-generation/recommended-workflow). Ready to edit? [Try the AI Image Editor](https://www.3daistudio.com/ImageGeneration/edit) -- free credits on signup. ## Frequently Asked Questions ### Can I edit a photo using just a text description? Yes. In 3D AI Studio, you upload any image and type what you want changed in plain English. For example, "change the background to a beach at sunset" or "make this person look like a cartoon character." The AI interprets your text and applies the edit while preserving parts you did not mention. ### What is the difference between GPT Image 2 Edit and Flux 2 Edit? GPT Image 2 Edit (3 credits) excels at precise, targeted changes like swapping specific objects, adding text, and photorealistic adjustments. Flux 2 Edit (4 credits) is better for style transformations and artistic changes across the whole image. In 3D AI Studio, you can try both on the same image to see which produces better results for your specific edit. ### How do I replace a background in an image with AI? Upload your image to the Edit tool in 3D AI Studio and write a prompt like "replace the background with a cozy coffee shop interior" or "make the background a plain white studio." The AI keeps your subject intact and generates a new background. For cleanest results, use GPT Image 2 Edit or Nano Banana 2 Edit. ### Can I turn a photo into a cartoon or illustration style? Absolutely. Upload your photo and write something like "turn this into a Pixar-style 3D character" or "make this look like anime art." 3D AI Studio's edit models handle full style transfers while preserving the person's likeness and pose. Nano Banana 2 Edit and GPT Image 2 Edit both work well for this. ### Does AI editing reduce image quality or resolution? The output resolution depends on the model and settings you choose. GPT Image 2 Edit outputs at fixed sizes, while Nano Banana 2 Edit supports flexible resolution from 0.5K to 4K. In 3D AI Studio, you can upscale the edited result afterward using the Upscale tools if you need a higher resolution version. ### How much does AI image editing cost? In 3D AI Studio, editing starts at 3 credits per edit with GPT Image 2 Edit. Nano Banana 2 Edit ranges from 4-12 credits depending on resolution, and Nano Banana Pro Edit costs 8-14 credits for the highest quality. Most users find GPT Image 2 Edit offers the best value for typical edits. ### Can I use multiple reference images when editing? Yes. Nano Banana 2 Edit supports up to 14 reference images, and Flux Kontext lets you combine and edit up to 4 images together. This is useful in 3D AI Studio when you want to blend elements from multiple sources or maintain style consistency across a series of edits. ### What types of edits work best with AI? Style transfers, background changes, and object swaps produce the most reliable results. Edits that work less well include very precise pixel-level adjustments, changing small text, or modifying hands and fingers. In 3D AI Studio, the key to good results is being specific about what to change and what to keep in your prompt. ### How do I prepare an edited image for 3D model conversion? After editing in 3D AI Studio, clean up the background (make it white or transparent), center the subject, and fix any AI artifacts. Use flat, even lighting and make sure the object fills the frame. Then send it to Image to 3D directly from the editor with one click. ### Can I undo an AI edit or go back to the original? 3D AI Studio keeps your original image intact. Each edit creates a new output, so your source is never modified. You can iterate on the same original with different prompts as many times as you want, comparing results side by side until you get the perfect edit. --- # Convert Source: https://docs.3daistudio.com/image-studio/convert Convert transforms your images into completely different art styles. Upload any image, pick a conversion style, and the AI recreates it in that style while keeping the core subject and composition intact. It is a fast way to explore how your concepts look as figurines, anime characters, oil paintings, pixel art, and more. ![Bust conversion example](https://docs.3daistudio.com/assets/images/docs/is-convert-bust.jpg) *A photo converted into a classical bust sculpture style.* ![Figurine conversion example](https://docs.3daistudio.com/assets/images/docs/is-convert-figurine.jpg) *A photo converted into a tabletop figurine style.* ## Available Styles Convert includes 14 built-in styles. Some are available on the free plan, and some require a paid plan. | Style | Description | Plan | | --- | --- | --- | | Convert Anything | General-purpose style transfer. Describe the target style in your prompt. | Free | | Parts Sheet | Breaks a character or object into individual component parts, laid out as a reference sheet. | Paid | | Bust Maker | Transforms portraits into classical sculptural busts with realistic stone or bronze materials. | Paid | | Table-top Figurine | Converts characters into detailed tabletop miniature figurines. | Free | | Pose Generator | Changes a character's pose to T-pose, A-pose, or any custom pose you describe. | Free | | Fantasy Character | Transforms images into high-fantasy character art with armor, weapons, and magical elements. | Free | | Stylized 3D Render | Recreates the image as a polished 3D render with stylized lighting and materials. | Free | | Anime Character | Converts any portrait or character into anime art style. | Free | | Ghibli Art | Applies the warm, painterly look of Studio Ghibli films. | Free | | Brick Figure | Turns characters into collectible brick-style toy figures. | Free | | Oil Painting | Applies traditional oil painting techniques with visible brushstrokes and rich color. | Paid | | Comic Panel | Converts the image into a comic book panel with bold lines, halftone shading, and speech bubbles. | Paid | | Pixel Art Sprite | Creates retro pixel art sprites from any image. Great for game dev assets. | Paid | | Claymation | Makes the image look like a stop-motion clay figure with soft, sculpted textures. | Paid | ## How It Works Using Convert is straightforward: 1. **Upload an image.** Any photo, AI-generated image, or artwork works. The clearer and better-lit the input, the better the result. 2. **Pick a conversion style** from the list above. 3. **Choose an AI model.** Convert supports three models: - **Nano Banana 2** (4-12 credits) for flexible resolution and fast results. - **Nano Banana Pro** (8-14 credits) for the highest quality output. - **GPT Image 2** (3 credits) for affordable, high-quality conversions. 4. **Add optional instructions.** You can guide the conversion with extra details in the prompt. For example, when using Bust Maker you could add "white marble material, dramatic side lighting" to control the look. 5. **Generate.** Hit `Ctrl+Enter` or click Generate. ## Tips ### Best input images - **Clear, well-lit photos** convert better than dark or blurry ones. The AI needs to see the subject clearly to recreate it in a new style. - **Simple backgrounds** help. A busy background can confuse the conversion, especially for styles like Figurine or Bust that need to isolate the subject. - **Front-facing or three-quarter views** tend to produce the best results for character-based styles like Anime, Fantasy, and Figurine. ### Best styles for 3D conversion If you plan to turn the converted image into a 3D model with [Image to 3D](https://www.3daistudio.com/ImageTo3D) ([guide](https://docs.3daistudio.com/3d-generation/image-to-3d)), some styles work better than others: - **Stylized 3D Render** is the most natural fit since the output already looks like a 3D model. - **Table-top Figurine** and **Bust Maker** both produce images with clear depth and volume that translate well to 3D. - **Brick Figure** creates clean, geometric shapes that 3D generators handle easily. - Avoid **Oil Painting** and **Comic Panel** for 3D conversion since the flat, stylized look can confuse depth estimation. Ready to convert? [Try the Image Style Converter](https://www.3daistudio.com/ImageGeneration/convert) -- free credits on signup. ## Frequently Asked Questions ### Can I convert a photo into anime style with AI? Yes. In 3D AI Studio, upload any photo to the Convert tool and select "Anime Character" from the style list. The AI transforms your image into anime art while keeping the original pose, composition, and key features. It works with portraits, full-body shots, and even group photos. ### How does AI style transfer work? AI style transfer analyzes your input image for structure, subject, and composition, then regenerates it in the target art style. In 3D AI Studio's Convert tool, you pick from 14 preset styles and the AI handles the rest. The "Convert Anything" option also lets you describe any custom style in your prompt. ### Can I turn a photo into an oil painting with AI? Yes. Select the "Oil Painting" style in 3D AI Studio's Convert tool. The AI applies traditional oil painting techniques with visible brushstrokes, rich color, and natural canvas texture. This style requires a paid plan and costs 3-14 credits depending on your chosen model and resolution. ### How do I turn a photo into a figurine or miniature? Upload your image and select "Table-top Figurine" in 3D AI Studio's Convert tool. The AI transforms your subject into a detailed tabletop miniature with painted details and a sculpted base. This style is free to use and works especially well as input for 3D model conversion afterward. ### What is the cheapest way to convert image styles? GPT Image 2 at 3 credits per conversion is the most affordable model in 3D AI Studio's Convert tool. It handles all 14 styles with good quality. For higher resolution or more detail, Nano Banana 2 (4-12 credits) and Nano Banana Pro (8-14 credits) are available as step-up options. ### Can I convert images to pixel art for game development? Yes. The "Pixel Art Sprite" style in 3D AI Studio creates retro-style pixel art from any image. It works great for creating game assets, character sprites, and icon sets. This style requires a paid plan and produces clean, game-ready output that you can use directly in pixel-based games or as reference. --- # Sketch to Image Source: https://docs.3daistudio.com/image-studio/sketch Turn your rough sketches and drawings into polished, detailed images using AI. Whether you upload an existing sketch or draw directly in the browser, the AI interprets your lines and produces a finished result based on your prompt. ![Sketch to Image tool interface](https://docs.3daistudio.com/assets/images/docs/is-sketch-to-image.jpg) *The Sketch to Image tool interface showing a sketch being converted into a polished image.* ## Two Tools Image Studio includes two sketch tools: 1. **Sketch to Image** - Upload a sketch or drawing from your device. The AI uses it as a guide to generate a polished image. 2. **Paint & Transform** - Draw directly in the browser using the built-in canvas. No need to upload anything. Sketch your idea, then let the AI transform it. Both tools work the same way once your sketch is ready: write a prompt describing what you want the final image to look like, pick a model, and generate. ## Available Models | Model | Credits | | --- | --- | | Nano Banana 2 | 4 | | Nano Banana Pro | 8 | | GPT Image 2 | 3 | ## How to Use 1. Upload a sketch (Sketch to Image) or draw one in the browser (Paint & Transform). 2. Write a prompt describing the desired output. For example: "a cozy cabin in the mountains, watercolor style." 3. Pick a model from the dropdown. 4. Click Generate. The AI will use your sketch as a structural guide and produce an image that matches your prompt. ## Tips - **Clean lines work best.** Simple, clear outlines give the AI a stronger structure to follow. - **Add shading for depth.** Even rough shading helps the AI understand light direction and volume. - **Describe the style in your prompt.** The sketch gives structure, but the prompt controls the look. Mention things like "oil painting", "anime style", or "photorealistic" to guide the output. - **Use bold shapes.** Thin, faint lines may get lost. Thicker strokes translate better. ## Sketch to 3D Workflow You can combine Sketch to Image with Image to 3D to go from a hand drawing to a full 3D model. See the [Sketch to 3D workflow guide](https://docs.3daistudio.com/3d-generation/recommended-workflow#workflow-3-sketch-to-3d) for a step-by-step walkthrough. Ready to try it? [Open the Sketch to Image tool](https://www.3daistudio.com/ImageGeneration/sketch) -- free credits on signup. ## Frequently Asked Questions ### Can AI turn a rough sketch into a realistic image? Yes. In 3D AI Studio, upload any sketch or doodle to the Sketch to Image tool, write a prompt describing the desired result, and the AI produces a polished, detailed image. Even very rough pencil sketches work - the AI uses your lines as structural guidance and fills in color, texture, lighting, and detail based on your prompt. ### What is the Paint and Transform tool? Paint and Transform is a built-in drawing canvas inside 3D AI Studio's Image Studio. You draw directly in your browser using brushes and shapes, then the AI transforms your sketch into a finished image. It saves you the step of drawing on paper, photographing it, and uploading. Both tools produce the same quality output. ### How detailed does my sketch need to be for good results? Simple outlines with basic shading are enough. 3D AI Studio's sketch tools work with anything from rough stick figures to detailed line art. Clean, bold lines give the best results because thin or faint strokes may get lost. Adding even rough shading helps the AI understand light direction and depth. ### Which AI model is best for sketch to image conversion? GPT Image 2 (3 credits) offers the best balance of quality and affordability in 3D AI Studio. For maximum detail and resolution, Nano Banana Pro (8 credits) produces the highest quality output. Nano Banana 2 (4 credits) is a good middle option for fast iterations when you are still refining your sketch. ### Can I go from a hand drawing to a 3D model? Yes, 3D AI Studio supports a full sketch-to-3D pipeline. First, convert your sketch into a polished image using Sketch to Image. Then send that image to Image to 3D to generate a full 3D model. The whole process takes a few minutes and requires no 3D modeling skills. ### Does the sketch tool work with tablet or stylus input? Yes. The Paint and Transform canvas in 3D AI Studio supports any pointing device including mouse, tablet, stylus, and touchscreen. Pressure sensitivity is not currently supported, but you can adjust brush size manually to simulate thick and thin strokes. --- # AI Video Generation Source: https://docs.3daistudio.com/image-studio/video Generate short AI videos from images or text prompts. Upload a starting image and describe the motion you want, or generate a video entirely from a text description. Multiple models are available with different strengths, quality levels, and pricing. ![Video generation in progress](https://docs.3daistudio.com/assets/images/docs/is-video-generating.jpg) *Video generation in progress with a progress indicator.* ![Video generation result](https://docs.3daistudio.com/assets/images/docs/is-video-result.jpg) *Completed video generation showing the result preview.* [Video: Example](https://docs.3daistudio.com/assets/images/docs/is-video-example.mp4) ## Available Models | Model | Type | Credits | Best for | | --- | --- | --- | --- | | Veo 3 Fast | Image to Video | 60 | Quick animation | | Veo 3 Fast | Text to Video | 60 | Text-only video | | Kling | Image to Video | 40-80 | Standard/Master quality | | Veo 3 Standard | Image to Video | 120 | Premium quality | | Lucy 14B | Image to Video | 60 | Creative animation | | Seedance 1.5 Pro | Image to Video | 80-200 | Advanced settings, end frame | | Kling O1 | Frame Transition | 80-200 | Start/end frame morph | | Kling Motion Control | Motion Transfer | 80-200 | Transfer motion from ref video | Credit costs vary based on quality settings and video duration where applicable. ## Video Types **Image to Video** - Upload a starting frame and describe how it should move. The AI animates your image into a short video clip. **Text to Video** - Describe a scene entirely in text. The AI generates both the visuals and the motion from scratch. **Frame Transition** - Provide a start frame and an end frame. The AI creates a smooth morph between them. **Motion Transfer** - Upload a reference video and a source image. The AI applies the motion from the reference video to your image. ## How to Use 1. Select a video model from the dropdown. 2. Upload an input image (or write a text prompt for Text to Video). 3. Write a motion prompt describing what should happen. Keep it simple - "camera slowly zooms in", "character turns head to the left", "leaves blowing in the wind." 4. Adjust quality settings if available for your chosen model. 5. Click Generate. Video generation takes longer than image generation - typically 30 seconds to a few minutes depending on the model. ## Tips - **Clear input images work best.** High-contrast subjects with simple backgrounds produce smoother animations. - **Keep motion prompts simple.** Describe one or two motions rather than complex choreography. "Camera pans left" works better than a paragraph of instructions. - **Use Veo 3 Fast for testing ideas.** Once you find a prompt and input that works, switch to a higher-quality model for the final render. - **Frame Transition is great for turntables.** Use a front view as the start frame and a side view as the end frame to create a smooth rotation. ## Requirements All video generation tools require a paid plan. Free accounts cannot access video features. [Try the AI Video Generator](https://www.3daistudio.com/ImageGeneration/video) ### Related Tools - Need a starting image? [Generate one with AI](https://docs.3daistudio.com/image-studio/generate) or [edit an existing photo](https://docs.3daistudio.com/image-studio/edit) - Want a 3D turntable video instead? Use the [Model Spinner](https://docs.3daistudio.com/utilities/model-spinner) for rotating 3D model videos - Want to turn images into 3D models? Try [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) ## Frequently Asked Questions ### How much does AI video generation cost? In 3D AI Studio, video generation ranges from 40 to 200 credits per video depending on the model and quality settings. Veo 3 Fast starts at 60 credits for a quick animation, while premium models like Seedance 1.5 Pro can cost up to 200 credits for longer, higher-quality output with advanced controls. ### What is the difference between Veo 3 and Kling for AI video? Veo 3 (by Google) produces the highest quality output with smooth motion and realistic physics, available in Fast (60 credits) and Standard (120 credits) tiers. Kling offers more creative flexibility with standard and master quality modes (40-80 credits), plus unique features like frame transitions and motion transfer. Both are available in 3D AI Studio. ### Can I generate a video from just a text description? Yes. 3D AI Studio offers Text to Video through Veo 3 Fast (60 credits). Describe a scene and motion in your prompt, and the AI generates both the visuals and animation from scratch. Image to Video typically produces more predictable results since the AI has a clear starting point. ### How long are AI-generated videos? Most AI videos generated in 3D AI Studio are 3-5 seconds long. The exact duration depends on the model, with some offering configurable length. Generation time is typically 30 seconds to a few minutes, which is longer than image generation due to the complexity of creating motion frames. ### What is AI motion transfer and how does it work? Motion transfer lets you upload a reference video and a static image, then the AI applies the movement from the video onto your image. In 3D AI Studio, Kling Motion Control handles this at 80-200 credits. It is great for making product shots or characters perform specific actions without filming new footage. ### Do I need a paid plan for AI video generation? Yes. All video generation tools in 3D AI Studio require a paid plan. Free accounts cannot access video features. This includes Image to Video, Text to Video, Frame Transition, and Motion Transfer across all models like Veo 3, Kling, Lucy, and Seedance. --- # Image Utilities Source: https://docs.3daistudio.com/image-studio/utilities Image Studio includes five utility tools for common image tasks like removing backgrounds, upscaling, cropping, and fixing lighting. These are especially useful for preparing images before 3D conversion. ![Remove Background tool](https://docs.3daistudio.com/assets/images/docs/is-remove-bg.jpg) *The Remove Background tool showing a before and after result.* ## Available Tools ### 1. Remove Background (1-3 credits) Remove the background from any image, leaving just the subject on a transparent canvas. Three providers are available: | Provider | Credits | Notes | | --- | --- | --- | | BiRefNet | 1 | Fast, good for simple subjects | | BiRefNet v2 | 2 | Better edge detection | | Pixelcut | 3 | Best quality, handles hair and fine details | ### 2. Upscale Creative (1 credit) Enlarge images 2x to 4x with AI-generated detail. An optional prompt and creativity and detail sliders let you guide how the AI fills in new pixels. Good for stylized or artistic images where you want the AI to add creative detail. ### 3. Upscale Faithful (1 credit) Enlarge images 2x to 4x while preserving existing detail as closely as possible. Best for photos or renders where you want a clean, faithful enlargement without added interpretation. ### 4. Crop Image (free - 0 credits) Crop and resize images directly in the browser. This runs entirely on your device with no AI involved, so it costs nothing. Use it to trim unwanted areas or adjust aspect ratios before using other tools. ### 5. Restore Lighting (2 credits) Remove unwanted shadows and normalize lighting across an image. Designed to prepare images for 3D conversion where consistent lighting produces better results. ## Tips - **Use Remove Background before Image to 3D.** A clean subject on a transparent background gives 3D generators much better input than a busy photo. - **Use Restore Lighting before 3D conversion.** Shadows baked into the source image can cause artifacts in the 3D model. Removing them first leads to cleaner geometry and textures. - **Combine tools in sequence.** A typical prep workflow: Remove Background, then Restore Lighting, then Upscale if the image is small, then send to Image to 3D. [Try the Image Utilities](https://www.3daistudio.com/ImageGeneration/utilities) -- background removal, upscaling, and more. ### Related Tools - After cleaning up your image, convert it to 3D with [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) - See the full [Recommended Workflow](https://docs.3daistudio.com/3d-generation/recommended-workflow) for preparing images before 3D conversion - Need to edit the image content itself? Use [AI Image Edit](https://docs.3daistudio.com/image-studio/edit) ## Frequently Asked Questions ### What is the best AI tool for removing backgrounds from images? 3D AI Studio offers three background removal providers. BiRefNet (1 credit) is fast and good for simple subjects. BiRefNet v2 (2 credits) has better edge detection. Pixelcut (3 credits) handles hair, fur, and fine details the best. For most images, BiRefNet v2 gives the best balance of quality and cost. ### Can AI upscale images without losing quality? Yes. 3D AI Studio includes two upscaling tools. Upscale Faithful (1 credit) enlarges images 2x to 4x while faithfully preserving existing detail, best for photos and renders. Upscale Creative (1 credit) enlarges 2x to 4x and adds AI-generated detail, better for stylized or artistic images where you want the AI to enhance the result. ### How should I prepare an image before converting it to a 3D model? The recommended prep workflow in 3D AI Studio is: first remove the background, then restore lighting to remove baked-in shadows, then upscale if the image is small. This gives 3D generators a clean, well-lit subject on a transparent background, which produces much better geometry and textures than a raw photo. ### Does the crop tool cost any credits? No. The Crop Image tool in 3D AI Studio is completely free at 0 credits. It runs entirely on your device with no AI processing involved. Use it to trim unwanted areas, adjust aspect ratios, or resize images before sending them to other tools that do cost credits. --- # Character Sheet Generator Source: https://docs.3daistudio.com/image-studio/character-sheet Upload a single character image and get a 4-view turnaround sheet showing front, left, right, and back views. Then split the grid into individual panels for use as 3D modeling reference, game assets, or animation prep. ![Character sheet output with four views](https://docs.3daistudio.com/assets/images/docs/is-character-sheet-output.jpg) *A completed 4-view character turnaround sheet showing front, left, right, and back views.* ## What It Does The Character Sheet Generator takes one image of a character and produces a grid showing the character from four angles. This is the standard format used in game development and animation for character turnarounds. The AI maintains the character's appearance, proportions, and details across all views, giving you a consistent reference sheet from a single input. ## Available Models | Model | Credits | Notes | | --- | --- | --- | | GPT Image 2 | 6 (medium quality) | Default, most consistent results | | Nano Banana 2 | 4 | Fast and affordable | | Gemini Edit | 4 | Good color accuracy | | Nano Banana Pro | 6-12 | Higher resolution options | ## Settings - **AI Model** - Choose which model generates your sheet. - **Quality** (GPT Image 2 only) - Low, medium, or high detail. - **Resolution** (Nano Banana Pro only) - Control output image size. - **Additional Details** - Text field for extra instructions. For example: "fantasy armor", "casual clothing", or "simple flat colors." ## Grid Splitter After generation, the built-in grid splitter lets you break the turnaround sheet into individual panels: 1. **Drag the split lines** to align with the boundaries between views. 2. **Relabel views** if needed (front, left, right, back). 3. **Export individual panels** as separate JPEG files. This saves you from having to manually crop the output in an image editor. ## Access Free users can generate character sheets to preview the results. A paid plan is required to download the output images. ## Great For - **Game development** - Character turnarounds are the standard reference format for 3D character artists. - **3D modeling reference** - Pin each view beside your viewport while sculpting or modeling. - **Animation prep** - Consistent multi-angle reference helps keep characters on-model across frames. - **Concept art** - Quickly explore how a character looks from all sides before committing to a full design. ### Related Tools - Use character sheet views as multi-view input for [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) with multi-image models for more accurate 3D characters - Need to rig and animate the resulting 3D character? Use [Rigging & Animation](https://docs.3daistudio.com/processing/rigging) - Want to change character poses first? Try [Pose Transfer](https://docs.3daistudio.com/image-studio/pose-transfer) ## Frequently Asked Questions ### What is an AI character turnaround sheet? A character turnaround sheet shows the same character from four angles: front, left, right, and back. It is the standard reference format used in game development and animation. In 3D AI Studio, the Character Sheet Generator creates this automatically from a single image, keeping appearance and proportions consistent across all views. ### How many credits does a character sheet cost to generate? In 3D AI Studio, character sheet generation costs 4-12 credits depending on the model. Nano Banana 2 and Gemini Edit cost 4 credits each for fast results. GPT Image 2 costs 6 credits at medium quality and delivers the most consistent multi-view output. Nano Banana Pro ranges from 6-12 credits for higher resolution options. ### Can I use AI character sheets as reference for 3D modeling? Yes, that is one of the primary use cases. After generating a turnaround sheet in 3D AI Studio, use the built-in grid splitter to export individual view panels. Pin each view beside your 3D viewport while sculpting. You can also feed the split views directly into multi-view Image to 3D models for AI-generated 3D characters. ### Do I need a paid plan to use the Character Sheet Generator? Free users can generate character sheets to preview the results and see all four views. However, 3D AI Studio requires a paid plan to download the output images. This lets you test the tool and check quality before committing to a subscription. --- # AI Pose Transfer Source: https://docs.3daistudio.com/image-studio/pose-transfer Upload a character or person image, choose a target pose, and the AI generates your subject in that new pose while keeping their appearance intact. Pick from 21 preset poses or upload a custom pose reference. ![Pose Transfer settings with source image and pose reference](https://docs.3daistudio.com/assets/images/docs/is-pose-settings.jpg) *The Pose Transfer interface showing the source image and pose reference library.* ![Pose Transfer result - pose applied to character](https://docs.3daistudio.com/assets/images/docs/is-pose-output.jpg) *The result with the new pose applied to the character.* ## How It Works 1. **Upload your image** - A photo or illustration of a character or person. 2. **Pick a pose** - Choose from the pose library or upload your own reference image showing the desired pose. 3. **Select a model** and click Generate. 4. **Compare results** using the before/after slider. The AI keeps the subject's face, clothing, body type, and details consistent while repositioning them into the target pose. ## Available Models | Model | Credits | Notes | | --- | --- | --- | | GPT Image 2 | 6 | Default, best consistency | | Nano Banana 2 | 4 | Fast and affordable | | Gemini Edit | 6 | Good detail preservation | | Nano Banana Pro | 8-16 | Higher resolution | ## Pose Library The built-in pose library includes 21 preset poses covering common positions like standing, sitting, action poses, and casual stances. You can favorite poses you use often for quick access. If the library does not have what you need, upload any image as a custom pose reference. The AI will extract the body position from your reference and apply it to your subject. ## Results Generated images include a before/after comparison slider so you can quickly see how the pose changed relative to the original. You can download the result or send it to other Image Studio tools for further editing. ## Access Free users can generate pose transfers to preview the results. A paid plan is required to download the output images. ## Great For - **Character design** - Explore how a character looks in different poses without redrawing. - **Social media content** - Create dynamic variations of product shots or portraits. - **Product photography** - Repose mannequins or models without reshooting. - **3D reference** - Generate specific poses as sculpting or modeling reference. ### Related Tools - Need multi-angle views of a character? Use the [Character Sheet Generator](https://docs.3daistudio.com/image-studio/character-sheet) - Ready to convert to 3D? Send your posed image to [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) - Want to rig and animate a 3D character? Use [Rigging & Animation](https://docs.3daistudio.com/processing/rigging) ## Frequently Asked Questions ### How does AI pose transfer work? AI pose transfer takes your original character image and a target pose reference, then generates the character in that new position. The AI in 3D AI Studio preserves the subject's face, clothing, body type, and details while repositioning them. It uses advanced image understanding to figure out how each body part should look from the new angle. ### Can I use my own custom pose reference image? Yes. Beyond the 21 preset poses in 3D AI Studio's pose library, you can upload any image as a custom pose reference. The AI extracts the body position from your reference and applies it to your subject. A photo of a person, a stick figure, or even a screenshot from a video all work as pose references. ### How much does AI pose transfer cost? In 3D AI Studio, pose transfer costs 4-16 credits depending on the model. Nano Banana 2 is the cheapest at 4 credits. GPT Image 2 and Gemini Edit cost 6 credits each and offer the best consistency. Nano Banana Pro ranges from 8-16 credits for higher resolution output. ### What types of images work best for pose transfer? Clear, well-lit images of characters or people against simple backgrounds produce the best results in 3D AI Studio. Full-body shots work better than cropped images since the AI needs to see the whole figure to repose it. Both photos and illustrations work, though illustrated characters with clean outlines tend to maintain consistency better across poses. --- # Image to Prompt Source: https://docs.3daistudio.com/image-studio/image-to-prompt Upload any image and get a detailed text prompt describing what is in it. Use the generated prompt to recreate similar images in other AI generators, learn how successful images are described, or save descriptions for reference. ![Image to Prompt settings panel](https://docs.3daistudio.com/assets/images/docs/is-image-to-prompt-settings.jpg) *The Image to Prompt settings panel showing style, length, and focus area controls.* ![Generated prompt output from image](https://docs.3daistudio.com/assets/images/docs/is-image-to-prompt-output.jpg) *The generated text prompt result from an uploaded image.* ## How It Works 1. Upload an image. 2. Configure your settings (style, length, language, focus areas). 3. Click Generate. 4. Copy the resulting text prompt with one click. The tool uses GPT-4o-mini to analyze your image and produce a text description tailored to your settings. There is no model picker - it always uses GPT-4o-mini. **Cost:** 2 credits per generation. ## Settings ### Description Style Choose how the prompt reads: - **Default** - Balanced description - **Detailed** - Thorough, covers everything visible - **Concise** - Short and focused - **Technical** - Emphasis on specs, measurements, rendering terms - **Artistic** - Emphasis on mood, composition, artistic qualities - **Keywords** - Comma-separated tags rather than sentences ### Prompt Length Slider from 10 to 450 words. Shorter prompts are punchier and work well for simple images. Longer prompts capture more nuance and are better for complex scenes. ### Output Language Available in 8 languages: English, German, French, Spanish, Italian, Portuguese, Japanese, and Chinese. ### Focus Areas Toggle up to 16 focus areas to guide what the AI pays attention to: colors, materials, lighting, composition, textures, mood, perspective, background, foreground, details, style, subject, environment, action, clothing, and anatomy. ## Access Image to Prompt requires a paid plan. It is not available on the free tier. ## Great For - **Reverse-engineering images** - See how an AI would describe a photo or artwork, then use that prompt to generate similar results. - **Learning prompt writing** - Study how detailed, effective prompts are structured by analyzing images you like. - **Recreating styles** - Get the prompt for a style you want to replicate, then use it across multiple generations. - **Documentation** - Save text descriptions of visual assets for your project files. ### Related Tools - Use extracted prompts to [Generate new images](https://docs.3daistudio.com/image-studio/generate) in a similar style - See [Prompt Tips](https://docs.3daistudio.com/3d-generation/prompt-tips) for how to refine and improve prompts for better results - Ready to convert images to 3D? Try [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) ## Frequently Asked Questions ### Can AI extract a text prompt from any image? Yes. Upload any photo, artwork, AI-generated image, or screenshot to the Image to Prompt tool in 3D AI Studio, and it generates a detailed text description. The tool uses GPT-4o-mini to analyze the image and produces a prompt you can copy with one click. It costs 2 credits per generation. ### How do I reverse-engineer an AI image's prompt? Upload the image to 3D AI Studio's Image to Prompt tool and select "Detailed" style with a high word count (200-450 words). The AI will describe the subject, composition, lighting, colors, style, and mood in detail. You can then paste this prompt into any AI image generator to recreate a similar result. ### What description styles are available for image-to-prompt? 3D AI Studio offers six description styles: Default (balanced), Detailed (thorough), Concise (short), Technical (specs and rendering terms), Artistic (mood and composition focus), and Keywords (comma-separated tags). You can also set the prompt length from 10 to 450 words and choose from 8 output languages. ### Is Image to Prompt free to use? Image to Prompt requires a paid plan in 3D AI Studio and costs 2 credits per generation. It is not available on the free tier. However, since each generation is only 2 credits, it is one of the most affordable tools available and you can analyze many images without spending significant credits. --- # Flow Overview Source: https://docs.3daistudio.com/flow/overview Flow is the visual node editor inside 3D AI Studio. Instead of running one tool at a time, you place each step as a "node" on an infinite canvas, draw wires between the nodes to pass data from one step to the next, and then run the whole pipeline with a single click. Think of it as the difference between cooking one dish at a time versus setting up a kitchen line. Once your line is set up, you can run the same recipe again and again, change one ingredient, and get a fresh result without rebuilding anything. Flow is part of [3D AI Studio](https://www.3daistudio.com), the all in one platform for turning text and images into 3D models. If you are new here, the standalone [Image to 3D generator](https://www.3daistudio.com/ImageTo3D) is the quickest way to see what the models can do before you start chaining them together in Flow. > [!NOTE] > Flow is currently in a private beta. The first time you open it you may see a "Join the beta" screen. Click the button to opt in and you will be taken straight into the editor. You can leave the beta anytime in Settings. ![A complete Flow pipeline on the canvas, wiring prompts, image generation, and 3D generation together](https://docs.3daistudio.com/assets/images/docs/flow/flow-full-example.webp) *A complete Flow pipeline on the canvas. Each card is a step, and the wires carry data from one step to the next.* ## Who Flow is for Flow works for two very different kinds of people, and it is built to serve both at once. - **Beginners** can start from a ready made template, change a prompt or an image, and press Run. You never have to understand the wiring to get a result. - **Power users** can build complex multi stage pipelines: generate a reference image, make four turnaround views, convert those into a 3D model, retexture it, and export it, all on one canvas that you can rerun forever. If you have ever used a node editor like Blender's shader nodes, Unreal's Blueprints, or ComfyUI, you will feel at home immediately. If you have not, do not worry. The rest of this guide explains every piece from the ground up. ## What you can do with Flow Flow connects almost every generation tool in 3D AI Studio into one canvas. In a single pipeline you can: - Write text prompts and combine them. - Generate images from text. - Edit images with AI using plain language instructions. - Use moodboards and style references to keep a consistent look. - Turn one image into four orthographic views (front, left, right, back). - Convert images into full 3D models. - Generate 3D models directly from text. - Retexture, retopologize (remesh), and convert the format of 3D models. - Compare results side by side and view them full screen. - Run an entire pipeline, queue several runs, and collect every output. You can also ask the built in AI assistant, named Bob, to build or edit the whole pipeline for you in plain English. ## The core ideas (read this first) Flow has a small number of concepts. Once these click, everything else makes sense. ### Nodes A **node** is a single step in your pipeline, drawn as a card on the canvas. Each node does one job: a "Generate Image" node turns a prompt into an image, an "Image to 3D" node turns an image into a model, a "Prompt" node holds text, and so on. Nodes are covered in full in the [Nodes Reference](https://docs.3daistudio.com/flow/nodes). ### Handles and wires Every node has small dots on its edges called **handles**. Handles on the left are **inputs** (data coming in), handles on the right are **outputs** (data going out). You connect an output handle to an input handle by dragging a **wire** between them. Wires are color coded by the kind of data they carry, so you can read a pipeline at a glance: | Color | Data type | Example | | --- | --- | --- | | Red | Text | A prompt | | Blue | Image | A single picture | | Indigo | Image set | A group of images (a moodboard) | | Green | Mesh | A 3D model | | Amber | Number | A seed or a setting value | | Violet | LLM | Text from a language model | | Cyan | Viewer | A preview passthrough | You can only connect handles of the **same type**. A blue image output will not connect to a red text input, which prevents you from building broken pipelines by accident. ### Running When you press **Run**, Flow works out the correct order to execute your nodes (everything a node depends on runs before that node), then runs them one at a time and feeds each result into the next step. Running is covered in full in [Running Flows](https://docs.3daistudio.com/flow/running-flows). ### Flows, templates, and runs - A **flow** (also called a canvas) is your editable document, the pipeline you are building. - A **template** is a saved snapshot of a flow that you or someone else can reuse. - A **run** is one execution of a flow with a specific set of inputs. Templates and sharing are covered in [Templates & Sharing](https://docs.3daistudio.com/flow/templates). ## A simple first pipeline Here is what a basic "text to 3D, with control" pipeline looks like in words: 1. A **Prompt** node holds the text "a cute wooden robot toy". 2. A wire carries that text into a **Generate Image** node, which produces a picture. 3. A wire carries that image into an **Image to 3D** node, which produces a model. 4. You press Run, and the three steps execute in order: prompt, then image, then model. The advantage over running the tools separately is that you can now change the prompt to "a cute metal robot toy", press Run again, and the whole chain regenerates. Nothing has to be re wired. The 3D step gives you the same quality as the standalone [Image to 3D tool](https://www.3daistudio.com/ImageTo3D), just inside a repeatable pipeline. ![A Flow pipeline that removes an image background and converts the result into a 3D model](https://docs.3daistudio.com/assets/images/docs/flow/flow-removebg-to-3d.webp) *A slightly longer pipeline: an image has its background removed, then the clean result is converted into a 3D model.* > [!NOTE] > **Why build the image first instead of going straight from text to 3D?** Because you get to see and perfect the picture before spending credits on a 3D model. Images are cheaper than models, so you can iterate freely on the look, then convert only once you are happy. ## How this guide is organized This documentation walks through the whole tool from start to finish. Read it in order for a complete understanding, or jump to the section you need. 1. **[The Flows Dashboard](https://docs.3daistudio.com/flow/dashboard)** - the home screen where you see your flows, create new ones, and browse templates. 2. **[The Canvas Editor](https://docs.3daistudio.com/flow/interface)** - the editor layout, toolbar, panels, navigation, and keyboard shortcuts. 3. **[Nodes Reference](https://docs.3daistudio.com/flow/nodes)** - how handles and connections work, plus every node explained. 4. **[Running Flows](https://docs.3daistudio.com/flow/running-flows)** - running single nodes and full pipelines, the run panel, the queue, and turning a flow into a reusable app. 5. **[AI Assistant (Bob)](https://docs.3daistudio.com/flow/ai-assistant)** - building and editing pipelines in plain English. 6. **[Templates & Sharing](https://docs.3daistudio.com/flow/templates)** - saving, sharing, publishing, and version history. 7. **[Models, Credits & Downloads](https://docs.3daistudio.com/flow/models-credits-downloads)** - the model catalog, parameters, credit costs, and downloading results. ## Frequently Asked Questions ### What is Flow in 3D AI Studio? Flow is a visual node editor where you build AI pipelines by connecting nodes on a canvas. Each node is one step (a prompt, an image generation, a 3D conversion, and so on), and wires pass data between them. You build the pipeline once and can rerun it with different inputs as many times as you like. ### Do I need to know node editors to use Flow? No. You can start from a template, change the prompt or image, and press Run without ever touching the wiring. Node editing is there when you want more control, but it is optional. Beginners can get results in minutes. ### How is Flow different from using the tools separately? The normal tools run one operation at a time. Flow lets you chain many operations into one repeatable pipeline. Change one input and rerun the whole chain. It is ideal when you do the same multi step process repeatedly or want consistent, reproducible results. ### Does Flow cost anything to use? Building and wiring pipelines costs nothing. You spend credits only when you run nodes that call AI models, exactly like the standalone tools. ### Is Flow similar to ComfyUI? Conceptually yes. Both are node based pipeline builders for AI generation. Flow is built into 3D AI Studio, runs entirely in your browser with no setup, focuses on the image and 3D generation tools in the platform, and includes an AI assistant that can build pipelines for you in plain English. [Open Flow](https://www.3daistudio.com/Flow) - [Try Image to 3D](https://www.3daistudio.com/ImageTo3D) - [3D AI Studio home](https://www.3daistudio.com) --- # The Flows Dashboard Source: https://docs.3daistudio.com/flow/dashboard The dashboard is the screen you land on when you open Flow. It is your home base: every flow you create lives here as a card, and this is where you start new flows, browse templates, and manage what you already have. ![The Flows dashboard showing the My Canvases tab with flow cards, tabs, and the New flow button](https://docs.3daistudio.com/assets/images/docs/flow/flow-dashboard-overview.webp) *The Flows dashboard. The tab bar runs across the top, and each flow appears as a card with its outputs as a preview.* ## Before the dashboard: the beta gate Flow is in a private beta, so the very first time you visit you may see a welcome gate instead of the dashboard. - If you are **signed in**, you will see a "Join the beta and enter" button. Click it once and you are in, with no reload and no waiting. You can leave the beta later in Settings under General. - If you are **signed out**, you will see a "Sign in to continue" button that takes you to the login screen and brings you back to Flow afterward. Once you have joined the beta, you go straight to the dashboard every time. ## Layout of the dashboard The dashboard is organized top to bottom: 1. **Explore banner** - a hero strip at the very top that points you toward the public template gallery. It changes its message depending on which tab you are on, and you can collapse it to a thin bar if you want more room. 2. **The tab bar** - a single row with a sidebar toggle on the left, the main tabs in the center, and a set of controls on the right that change per tab (filters, a page switcher, and the New flow button). 3. **The content grid** - the cards for whichever tab is active. On the left, the standard app **sidebar** slides in and out. Use the toggle button at the start of the tab bar to show or hide it. ## The tabs The tab bar has three tabs: | Tab | What it shows | | --- | --- | | **My Canvases** | Your own flows, the documents you are building. This is the default tab. | | **My Templates** | Templates you have saved or submitted, split into private drafts and public copies. | | **Public Gallery** | Templates published by the whole community that anyone can use. | Switching tabs is instant and just changes what the grid shows. ## My Canvases: your flows This is the main view and where you spend most of your dashboard time. ### What each card shows Every flow appears as a card. The card itself is the preview: it shows the flow's output images as a small stack of tilted tiles, with the flow's name, its icon, and how long ago it was last updated. Two extra markers can appear: - A green **Run** badge means the flow is set up as a runnable "app", meaning it has at least one promoted input (more on that in [Running Flows](https://docs.3daistudio.com/flow/running-flows)). - An amber **star** means the flow is favorited. ### Searching, sorting, and filtering The controls on the right side of the tab bar let you narrow the list, and they stay fast even when you have a large number of flows. - **Search** by name. Results update as you type (with a short debounce). - **Favorites filter** shows only flows you have starred. - **Sort** by most recent (the default, which also pins favorites to the top), by name (A to Z), or by oldest first. If there are more flows than fit on one page, a small page switcher appears showing the current page and the total. Flows are paged eight at a time. ### Loading and empty states - While the first page loads, you see skeleton placeholder cards. - If a search or filter returns nothing, you get a "No canvases match" message with a **Clear filters** button. - If you have no flows at all, you get a friendly "Create your first flow" screen with a **New flow** button and an **Explore the gallery** link. ## What you can do with a flow (card actions) Hover over a flow card and a row of action buttons appears in the corner. Each one does a specific thing: | Action | What it does | | --- | --- | | **Open** (click the card) | Opens the flow in the editor with a full screen transition. You can also press Enter or Space when a card is focused. | | **Share link** | Opens the share dialog and mints a short lived link to the flow. See [Templates & Sharing](https://docs.3daistudio.com/flow/templates). | | **Rename** | Opens a small dialog to change the flow's name and pick a new icon. | | **Duplicate** | Makes an exact copy of the flow. The copy appears at the top of the list since it is the newest. | | **Delete** | Removes the flow after a confirmation. | > [!NOTE] > Deleting a flow removes the canvas layout, but **your generated assets stay safe in your main Dashboard**. The 3D models and images you produced are not deleted, only the node pipeline that created them. ## Opening a flow Click a flow card and it opens in the editor with a smooth full screen transition that expands from the card. If a flow can no longer be opened (for example you deleted it), Flow lets you know and takes you back to the dashboard. ## Creating a new flow Click the **New flow** button (top right of the My Canvases tab) to open the launcher. The launcher offers four different ways to start, listed down the left side. ### 1. Blank canvas Start completely empty. Give the flow a name, pick an icon, and click **Create flow**. You land in an empty editor ready to build. This is the choice for when you know exactly what you want to make. ### 2. Build with AI Describe the pipeline you want in plain language, for example "generate a fantasy sword image, make four views of it, and turn it into a 3D model". Flow creates a fresh canvas, then hands your description to the AI assistant (Bob), who builds the whole pipeline automatically as soon as the editor opens. There are starter suggestions you can click, and you can submit with Ctrl+Enter or Cmd+Enter. > [!NOTE] > Build with AI uses credits, because the assistant is doing real work for you. The blank canvas and template paths do not cost anything to start. ### 3. From My Templates Pick one of your own saved templates as a starting point. Flow clones it into a brand new flow so your original template is untouched. ### 4. From Public Templates Start from a community template. The launcher shows a small set of official starter templates. There is also a link to browse the full gallery if you want more options. In every case, once the flow is created you fly straight into the editor with the same smooth transition used for opening an existing flow. ## My Templates tab This tab holds templates you personally own. It splits them into two groups with clear headers: - **Private drafts** - templates only you can see. You can keep editing these and submit them to the gallery later. - **Shared to the gallery** - public copies you submitted. Each shows its review state: In review, Live, or Not approved. Click any template to open its detail view, where all the management actions live (use, submit, share, delete). You can filter by visibility (all, drafts, shared) and search by name. See [Templates & Sharing](https://docs.3daistudio.com/flow/templates) for the full lifecycle. ## Public Gallery tab This is the community library of published templates. Browse, search by name, filter by category, and sort by: - **New** - most recently published. - **Trending** - gaining popularity right now. - **Most Used** - the most cloned templates. - **Official** - made by the 3D AI Studio team. - **Staff Picks** - hand selected highlights. Hover a template to preview its description, and click the heart to favorite it. Click **Use this template** to clone it into a new flow of your own. ## Deep links and reloading When you open a flow, the address bar updates so you can bookmark it or reload the page and land right back where you were. Bookmarks only open for flows you own or have been given access to. ## Frequently Asked Questions ### How do I create a new flow? Open Flow, make sure you are on the My Canvases tab, and click the New flow button in the top right. Choose one of four starting points: a blank canvas, Build with AI (describe it and the assistant builds it), one of your own templates, or a public template from the community. ### Where did my generated models go after I deleted a flow? They are still safe. Deleting a flow only removes the node pipeline (the canvas layout). The actual images and 3D models you generated stay in your main 3D AI Studio Dashboard and can be downloaded from there. ### Can I duplicate a flow? Yes. Hover over the flow card on the My Canvases tab and click the duplicate icon. Flow creates an exact copy that appears at the top of your list. This is useful when you want to try a variation without changing your original. ### What does the green "Run" badge on a flow card mean? It means the flow is set up as a runnable app, which happens when the flow has at least one promoted input (an input exposed in the run panel as a form field). These flows are designed to be run with simple inputs without editing the node graph. See Running Flows for details. ### How do I find a flow when I have a lot of them? Use the controls on the right of the My Canvases tab bar. Search by name, toggle the favorites filter to show only starred flows, or change the sort order (recent, name, or oldest). The list is paged eight flows at a time with a page switcher. ### Can I share a flow with someone who is not on my team? Yes. Hover the flow card and click the share icon to create a short lived link (valid for 10 minutes, 1 hour, or 24 hours). You can choose whether recipients can only view it or also clone it. See Templates and Sharing for the full details. [Open Flow](https://www.3daistudio.com/Flow) - [Try Image to 3D](https://www.3daistudio.com/ImageTo3D) - [3D AI Studio home](https://www.3daistudio.com) --- # The Canvas Editor Source: https://docs.3daistudio.com/flow/interface The canvas editor is where you build pipelines. It is an infinite canvas you can pan and zoom around, with your nodes placed on it and wires connecting them. This page is a complete tour of the interface: the chrome around the edges, how to add and connect nodes, how to organize the canvas, and every keyboard shortcut. ## The big picture Everything floats over the canvas in clearly defined zones, so the controls never get in the way of your work: - **Top left** tells you where you are: the flow name, a save status indicator, and a menu. - **Top center** is the editing toolbar: tools that change the graph. - **Top right** is the outcome cluster: the Run button, the AI assistant, and view toggles. - **Bottom left** holds flow level actions like saving a template and version history. - The **center** is the canvas itself. You can hide all of this chrome with the Hide UI toggle if you want a clean, distraction free view, and bring it back the same way. ## Top left: where am I This corner is about orientation and safety. - A **sidebar toggle** and **back button** to leave the flow. - The **flow name**, which you can rename inline by clicking it. - A **File menu** for actions like opening another flow, duplicating, or favoriting. - A **save status chip** that shows a checkmark when saved, a spinner while saving, a dot when there are unsaved changes, and an amber warning if a save failed. You can click it to force a save. > [!NOTE] > Flow **auto saves** your work in the background as you go. You rarely need to think about saving. The status chip is there so you always know the current state, and you can press Ctrl+S (or Cmd+S) to save immediately at any time. ## Top center: the editing toolbar This toolbar holds only the tools that change your graph, left to right: | Tool | What it does | | --- | --- | | **Add** | Opens the full categorized menu of nodes you can add. | | **Undo** | Reverses your last change. | | **Redo** | Reapplies a change you undid. | | **Drag / Select mode** | Switches between panning the canvas and rubber band selecting nodes. | | **Auto arrange** | Tidies up the layout, either of a selected section or the whole canvas. | | **Resolution toggle** | Switches image previews between fast (lightweight) and sharp (full resolution). | ### Drag mode versus Select mode This is an important distinction: - In **Drag mode** (the default), dragging the canvas with the left mouse button pans the view. To select nodes, click them or drag with the right or middle mouse button. - In **Select mode**, dragging with the left button draws a selection box (a "rubber band") around nodes. Panning moves to the right or middle mouse button. Press `S` to toggle Select mode and `Esc` to return to Drag mode. After you finish a rubber band selection, Flow automatically switches back to Drag mode so you can immediately move the group you just selected. ## Top right: outcomes and views This cluster is about results and how you look at them: - **Run** - the primary button that runs your flow. It has three states: an idle "Run" with a play icon, a live progress display with an inline Stop while running, and a selected state when the run panel is open. See [Running Flows](https://docs.3daistudio.com/flow/running-flows). - **AI Assistant (Bob)** - opens the assistant chat. See [AI Assistant](https://docs.3daistudio.com/flow/ai-assistant). - **Focus** - frames your whole flow neatly in view (also the `F` key). - **Outputs** - opens a drawer listing everything this flow has ever generated. - **Theme** and **Sound** toggles. - **Credits** - your current credit balance. - **Hide UI** - hides all the chrome for a clean view. ## Bottom left: flow level actions These actions apply to the whole flow rather than individual nodes: - **Save as template** - turn the current flow into a reusable template. - **Give feedback** - send feedback about the tool. - **Version history** - browse and restore earlier saved versions. - **Share link** - create a shareable link to the flow. - **Canvas settings**. ## Adding nodes (three ways) There are three ways to add a node, each suited to a different moment. ### 1. The Add menu (browse everything) Click **Add** in the toolbar, or right click anywhere on empty canvas, or press **Space** over empty canvas. This opens a categorized menu with every node, grouped into Text, LLM, Image, 3D (Mesh), Viewers, Logic, Workflows, and Notes. It also lists your custom nodes and any built in starter sub graphs. Use this when you want to browse what is available. ![The categorized Add node menu open on the canvas](https://docs.3daistudio.com/assets/images/docs/flow/flow-node-menu.webp) *The Add node menu, with nodes grouped by category.* ### 2. Quick Add (fast search) Double click empty canvas, or press the **`/`** key, to open a small searchable palette right at your cursor. Type a few letters to filter, use the arrow keys to move, and press Enter to add. This is the fastest way to drop a node when you know its name. ![The Quick Add search palette open at the cursor](https://docs.3daistudio.com/assets/images/docs/flow/flow-quick-add-search.webp) *Quick Add: press the slash key and type to find any node fast.* ### 3. Drag from a handle (add and connect at once) Grab an output handle on an existing node and drag the wire out into empty space. When you release, Quick Add opens already filtered to only the nodes that can accept that kind of data, and the new node is wired up automatically. This is the smoothest way to extend a pipeline. ![Dragging a wire out from a node output handle into empty space to add a connected node](https://docs.3daistudio.com/assets/images/docs/flow/flow-drag-from-handle.webp) *Drag a wire out from an output handle and release on empty canvas to add a compatible, pre wired node.* There is also a **Node Library** (opened from the AI assistant) that acts as a reference catalog. It shows each node's inputs and outputs, when to use it, gotchas, and the models it offers with their credit costs. > [!NOTE] > You can also **drag and drop files** straight onto the canvas. Drop an image to create an Image node, a 3D file to create a 3D Model node, and so on. And you can paste an image from your clipboard with Ctrl+V. ## Connecting nodes with wires To connect two nodes, drag from an output handle (right side, source) to an input handle (left side, target). A few rules and conveniences: - **Types must match.** A blue image output only connects to a blue image input. This prevents broken pipelines. - **Hollow versus filled dots.** An input that is optional and not yet connected shows as a hollow ring. Required or connected inputs are filled in. - **Single versus multiple inputs.** Some inputs accept only one wire (connecting a new one replaces the old). Others accept several, shown by progressively revealing more input slots as you connect. - **Click a connected dot to disconnect it.** You can also hover the middle of a wire and click the small "x" to remove it. - Wires are color coded by data type and animate when data flows during a run. See the [Nodes Reference](https://docs.3daistudio.com/flow/nodes) for the full list of data types and how each node's handles work. ## Organizing the canvas with sections A **section** (also called a frame) is a labeled box you can put nodes inside, similar to a frame in a design tool. Sections are the only nodes that can contain other nodes. - **Create a section from a selection:** select two or more nodes, then click "Create section" in the selection toolbar that appears above them. Flow draws a frame around them with a nice morph animation. - **Move nodes in and out:** drag a node onto a section to put it inside; drag it out to remove it. Dragging the section itself moves everything inside it together. - **Rename and recolor** sections from their label bar. - **Deleting a section keeps its children.** Removing a frame detaches the nodes inside rather than deleting them, so you never lose work by accident. Sections are the main tool for keeping a large pipeline readable. ## Selecting nodes - **Click** a node to select it. **Shift click** or rubber band (in Select mode) to select several. - When two or more nodes are selected, a small **selection toolbar** floats above them with quick actions: Create section, Save group (capture them as a reusable custom node), Clear outputs, Tag (apply a color tag), and Delete. - **Color tags** let you mark related nodes. A filter bar at the bottom lets you click a tag to highlight only nodes with that tag and dim the rest. ## The floating panels Four large panels can dock to the edges of the screen or be torn off to float freely: - **AI Assistant** - the Bob chat. - **Outputs** - the gallery of everything this flow generated. - **Run** - the run panel with inputs and the queue. - **Content Viewer** - a live preview that mirrors whatever node you have selected, shown larger. Drag a panel's header to move it. Drag it near an edge to dock it; drag it away to float it. Each panel remembers its size and position per flow. You can have only one panel docked to each edge at a time. ## Navigation and viewing - **Pan:** drag the canvas (in Drag mode) or use the right or middle mouse button. - **Zoom:** scroll wheel. Flow lets you zoom far out to see a huge pipeline at a glance. - **Focus your flow:** press `F` or click the Focus button to frame everything neatly. - **Fly to a node:** double click a node to zoom the camera to fit it. - **Level of detail:** when you zoom far out, heavy node interiors stop rendering to keep things smooth. Zoom back in and they return. - **Preview quality:** the resolution toggle in the toolbar switches images between a fast lightweight thumbnail and the full sharp asset. Wires, downloads, and large viewers always use the full asset regardless of this setting. ## Keyboard shortcuts These work when the canvas has focus (not while typing in a text field). | Shortcut | Action | | --- | --- | | `Space` (over empty canvas) | Open the Add node menu at the cursor | | `/` | Open Quick Add at the cursor | | Double click empty canvas | Open Quick Add | | `Ctrl+Z` / `Cmd+Z` | Undo | | `Ctrl+Shift+Z` / `Ctrl+Y` | Redo | | `Ctrl+S` / `Cmd+S` | Save now | | `Ctrl+C` / `Ctrl+V` | Copy / paste selected nodes | | `S` | Toggle Select mode | | `Esc` | Return to Drag mode | | `F` | Focus and frame the whole flow | | `Delete` / `Backspace` | Delete selected nodes or wires | | Double click a node | Fly the camera to that node | Quick add shortcuts for common nodes (press over empty canvas): | Key | Adds | | --- | --- | | `T` | Prompt (text) node | | `L` | LLM node | | `G` | Generate Image node | | `E` | Edit Image node | | `3` | Image to 3D node | | `C` | Combine Prompts (or collapse selected nodes if any are selected) | | `B` | Bypass (mute) selected generation nodes | ## Frequently Asked Questions ### How do I add a node in Flow? There are three ways. Click the Add button in the toolbar (or right click empty canvas, or press Space) to browse the full categorized menu. Press the slash key or double click empty canvas to open a quick search palette. Or drag a wire out from a node's output handle into empty space to add a connected node automatically. ### Why will my nodes not connect? Flow only lets you connect handles of the same data type. A blue image output cannot connect to a red text input, for example. Check that the output and input are the same color. Optional inputs that are not yet connected appear as hollow rings; required ones are filled dots. ### How do I save my work in Flow? Flow auto saves in the background as you work, so you usually do not need to do anything. The save status chip in the top left shows the current state. To save immediately, press Ctrl+S or Cmd+S, or click the chip. ### What is a section and how do I make one? A section is a labeled frame that groups nodes together, like a frame in a design tool. Select two or more nodes, then click Create section in the toolbar that appears above the selection. Dragging the section moves everything inside it. Deleting a section keeps the nodes inside, it only removes the frame. ### How do I pan and zoom the canvas? In the default Drag mode, drag with the left mouse button to pan and use the scroll wheel to zoom. Press F to frame your whole flow neatly. Double click a node to fly the camera to it. Switch to Select mode (press S) when you want left drag to draw a selection box instead. ### Can I copy nodes from one part of the canvas to another? Yes. Select the nodes and press Ctrl+C to copy, then Ctrl+V to paste. The pasted nodes keep the wiring between them. You can also duplicate a single node from its header menu. [Open Flow](https://www.3daistudio.com/Flow) - [Try Image to 3D](https://www.3daistudio.com/ImageTo3D) - [3D AI Studio home](https://www.3daistudio.com) --- # Nodes Reference Source: https://docs.3daistudio.com/flow/nodes Nodes are the building blocks of every pipeline. Each node is a card on the canvas that performs one job. This page first explains the system that all nodes share (handles, data types, and common controls), then documents every node grouped by category. ## How handles and data types work Every node has **handles**, the small dots on its edges. Inputs sit on the left, outputs on the right. You pass data from one node to another by wiring an output handle to an input handle. Each handle has a **data type**, and you can only connect handles of the same type. This is the single most important rule in Flow, and it is what stops you from building broken pipelines. The types are color coded: | Color | Type | Carries | | --- | --- | --- | | Red | `text` | A prompt or any string of text | | Blue | `image` | A single image | | Indigo | `imageset` | An ordered group of images (from a moodboard, for example) | | Green | `mesh` | A 3D model | | Amber | `number` | A numeric value such as a seed | | Violet | `llm` | Text produced by a language model | | Cyan | `viewers` | A passthrough used by viewer nodes | A few handle behaviors to know: - **Hollow ring** means an optional input that is not connected yet. **Filled dot** means it is required or already connected. - **Single input** handles accept one wire only; connecting a new wire replaces the old one. **Multiple input** handles reveal extra slots as you connect more wires. - **Click a connected dot** to disconnect that wire. ## Controls every node shares No matter what a node does, its header and body give you a consistent set of controls: - **Title and icon** identifying the node. - **Duplicate** to clone the node with its settings. - **Delete** to remove it. - **Tag** to apply a color tag for organization. - **Collapse** to shrink the node to a compact card (its handles stay so wires remain connected). - **Bypass (mute)** on generation nodes, which makes the node pass its input straight through without running. Press `B` with a node selected to toggle this. - During a run, a colored ring and a small number badge show the node's status and its position in the execution order. Generation nodes also keep their **result built into the node** (rather than spawning a separate node), so your downstream wiring stays attached even when you regenerate. They show a **history filmstrip** of previous results that you can click through. ![A 3D generation node showing its model picker, settings, result preview, and run bar](https://docs.3daistudio.com/assets/images/docs/flow/flow-3d-generation-node.webp) *A generation node with its model picker at the top, the generated result built into the node, and the run bar at the bottom. Input handles sit on the left, the output handle on the right.* ## Input and source nodes These nodes bring data into your pipeline. They are starting points, so they have nothing to wire in and instead produce an output you wire onward. ### Prompt (Text Input) The basic text box. Type your prompt and it flows out as `text`. You can resize it by dragging the corner. If you write something reusable, you can save it as a "pre prompt" custom node. - **Output:** text ### Image An image source or result. When empty it acts as a dropzone where you can upload a file or pick one from your Dashboard. When filled it displays the picture with download and history controls. Generation results often pop out into one of these. - **Output:** image ### Images (multi image) Holds up to ten images in one node, each in its own slot with its own output handle. This lets you wire several individual pictures downstream independently. Great when you want one node to feed many images into different places. - **Output:** one image handle per slot ![The Images node holding several pictures, each with its own output handle](https://docs.3daistudio.com/assets/images/docs/flow/flow-multi-image-node.webp) *The Images node holds several pictures in one place, each with its own output so you can wire them downstream independently.* ### Moodboard A pinboard of up to ten reference images that drives **style**. Its main output is a single ordered **image set** wire you connect to a Style Reference or Edit node. You can drag to reorder the pinned images, and there is a one click "Add Style Reference and connect" button that wires up a style edit for you. - **Inputs:** progressive image slots - **Output:** image set (plus optional per image handles) ![The Moodboard node showing pinned reference images that drive style](https://docs.3daistudio.com/assets/images/docs/flow/flow-moodboard-node.webp) *The Moodboard node collects reference images and outputs them as one ordered image set to drive a consistent style.* ### 3D Model A 3D source or result. When empty it lets you drop a model file (GLB, GLTF, STL, FBX, OBJ) or pick one from your Dashboard. When filled it shows a live, rotatable 3D preview with download and history. - **Output:** mesh ### Model Upload and Image Upload Minimal dropzones for adding a 3D file or an image into the pipeline. ### File A terminal node that surfaces a downloadable file, such as an export. It takes an input and gives you a download button. It has no output (it is an end point). - **Input:** generated result ## Text operator nodes These transform or produce text instantly and without using credits (except the LLM node). ### Combine Prompts Joins up to ten incoming prompts into one, in top to bottom order, with a separator you choose (comma, space, or new line). Instant, no credits. Useful for assembling a prompt from reusable pieces. - **Inputs:** up to ten text prompts - **Output:** text ### LLM Sends a prompt (and optionally up to six images) to a language model and outputs the resulting text. You pick a tier (Fast, Balanced, or Smart). Use it to write or refine prompts inside your pipeline, or to reason over images. - **Inputs:** prompt, optional images - **Output:** text ### Pre Prompt (custom) A reusable text operator that attaches a fixed prefix or suffix to whatever prompt you wire in. You create these from a Prompt node. Handy for always prepending a style instruction. ## Image generation nodes These call AI models to create or edit images. They all work the same way: pick a model, wire in its inputs, and the result appears right in the node, with a history of previous results and per node settings. ### Generate Image Turns text into an image. Pick from the curated image models and write a prompt. - **Input:** prompt (plus any settings you expose as handles) - **Output:** image ### Edit Image Edits one or more images using a text instruction, for example "make the background white" or "add a hat". Accepts individual images or a whole image set from a moodboard. - **Inputs:** image(s) or an image set, plus a prompt - **Output:** image ### Style Reference A specialized edit that generates something new in the visual style of the reference images you give it. It is tuned for style transfer, so you only need to describe what to generate. - **Inputs:** style reference images, prompt - **Output:** image ### Remove Background Cuts out the subject and produces a transparent image. No prompt needed. - **Input:** image - **Output:** image (transparent) ### Pose Changer and Optimize Image (curated one click edits) These are simplified, opinionated edit nodes locked to a single high quality edit model. Instead of writing a prompt, you flip simple toggles: - **Pose Changer** repositions a character to a T pose, A pose, or neutral pose while keeping its identity. - **Optimize Image** cleans an image up with toggles for centering, facing front, white background, removing clutter, and sharpening. ![The Pose Changer node with pose options](https://docs.3daistudio.com/assets/images/docs/flow/flow-pose-changer-node.webp) *The Pose Changer node. Pick a target pose and it repositions the character while keeping its identity.* ![The Optimize Image node with cleanup toggles](https://docs.3daistudio.com/assets/images/docs/flow/flow-optimize-image-node.webp) *The Optimize Image node. Flip simple toggles instead of writing a prompt.* ## 3D generation and derivation nodes These create or transform 3D models. ### Image to 3D, Multi View to 3D, Text to 3D Three nodes, one per input mode, each scoped to the matching model category so the default model and drag to connect work on the first try. These use the same models as the standalone [Image to 3D generator](https://www.3daistudio.com/ImageTo3D): - **Image to 3D** turns a single image into a model. - **Multi View to 3D** takes front, left, right, and back views (its handles line up exactly with the Multi View and Render Views nodes). - **Text to 3D** generates a model straight from a prompt. All three output a `mesh`. ### Texture 3D, Retopology (Remesh), Convert Format These operate on an existing 3D model: - **Texture 3D** applies a new texture to a model. - **Retopology (Remesh)** rebuilds the mesh with cleaner topology. - **Convert Format** changes the file format. Each takes a `mesh` input (plus an optional prompt or image depending on the model) and outputs a `mesh`. ## Image processing and transform nodes These reshape image data, mostly without generating anything new. ### Multi View Takes one image and produces four orthographic views (front, left, right, back), each as a separate output. When all four are ready it offers a one click "Create 3D model" button that wires them into a Multi View to 3D node. - **Input:** image - **Outputs:** front, left, right, back images ### Render Views (experimental) The inverse of Multi View: wire in a 3D model and it renders turnaround images locally, with **no credits used**. You choose the capture style (textured, unlit, clay, normals) and resolution. - **Input:** mesh - **Outputs:** an image set plus per view images ![The Render Views node rendering turnaround images from a 3D model](https://docs.3daistudio.com/assets/images/docs/flow/flow-render-views.webp) *Render Views takes a 3D model and renders turnaround images locally, with no credits used.* ### Combine to Grid and Split Grid A pair for consistent multi image editing: - **Combine to Grid** tiles several images into one contact sheet so an Edit node can edit them together, and records a layout for splitting them back. - **Split Grid** crops that edited grid back into individual images using the recorded layout. Combine to Grid even has a "Set up consistent edit" button that wires up the whole Combine, Edit, Split chain for you. ![A workflow rendering views of a 3D model, combining them to a grid, editing, then splitting back](https://docs.3daistudio.com/assets/images/docs/flow/flow-render-views-grid-workflow.webp) *A consistent edit workflow: Render Views feeds a Combine to Grid, which is edited as one image, then Split Grid breaks it back into individual pictures.* ### Split into Parts Drop in a character image and this node detects its parts automatically, shows you an editable checklist, then isolates each part as its own image. Each part gets both an image output and a text output (the extraction prompt). - **Input:** image - **Outputs:** per part images, an "all parts" image set ![The Split into Parts node detecting and isolating parts of a character image](https://docs.3daistudio.com/assets/images/docs/flow/flow-split-into-parts-node.webp) *Split into Parts detects the pieces of a character, then isolates each one as its own image.* ### Collect Results Taps a generator's history filmstrip and republishes its results as separate outputs, so every variation or batch result can flow downstream, not just the one active result on the generator's own output. It generates nothing itself. - **Input:** one image or one mesh source - **Outputs:** an **All results** image set (2D only) plus fixed numbered slots out1...outN - Pick the format (2D images or 3D meshes) and how many slots you want (1 to 10); the newest results fill the slots live. This is the key bridge for **chaining batches**. A batched generator only puts its latest result on its output, so to pass *all* of its results into another Batch, add a Collect after it and wire from its **All results** image set (not the numbered slots, which are capped by the slot count). See [Batch](https://docs.3daistudio.com/flow/nodes#batch). ## Logic and control nodes These are pure operators that route or hold values. They never call a model and cost no credits. ### Switch Routes exactly one of several same type inputs to its output. It has two modes: **manual** (you click a row to pick the active input) and **auto** (the first wired input that has content wins). The active branch is drawn inside the node. When a flow runs, only the selected branch executes, so a Switch is a real fork in your pipeline. - **Inputs:** several of the same type - **Output:** that type ### Seed Holds a random seed and outputs it as a number, so generators stay reproducible. You can set it to stay fixed or roll a new value each run. - **Output:** number ### Variable A named, typed value: text, number (with a slider), or a choice from a list. Variables are designed to be **promoted** into the run panel as a form field, turning your flow into a simple app. See [Running Flows](https://docs.3daistudio.com/flow/running-flows). - **Output:** text or number ### Batch Runs the node it feeds **once per item**, a "for each" over a collection of images, prompts, or 3D models. The Batch node is passive: it holds the collection and never runs itself; the generation or derivation node wired to its output is what repeats. - **Input:** several images or an image set, several prompts, or several 3D models (typed by the chosen mode) - **Output:** the same type, as a collection How to use it: - **Pick a type** (Images, Prompts, or 3D) with the toggle at the top. - **Fill it** either way: - **Manually** by dragging, dropping, or pasting images, or typing one prompt per line. - **By wiring** several sources into it (image or 3D nodes, a Moodboard, or a Collect's All results); the items are then derived live from those connections. - **Remove an item** by hovering it and clicking the **x** (for a wired item this disconnects that source). - Wire the Batch output into **one** generation or derivation node. That node runs once per item, and every result lands in its history filmstrip. Add a [Collect Results](https://docs.3daistudio.com/flow/nodes#collect-results) after it to fan all the results out to the next step. - **Test on one** runs only the first item, so you can dial in settings before committing to the whole batch. Supported on Generate Image, Edit Image, Style Reference, Remove Background, Image to 3D, Text to 3D (prompts), and Remesh, Texture, and Convert (3D). Not supported on Multi View to 3D, Multi View, or the AI (LLM) node. For several **variations of one prompt** you do not need a Batch; use the model's image count setting instead. Use a Batch when you have **many distinct inputs**. See [Batch runs](https://docs.3daistudio.com/flow/running-flows#batch-runs) for how a batch runs and what it costs. ## Viewer and output nodes These display results. They are end points and do not generate. ### Viewer A free size viewer for an image, a 3D model, or text. Wire in any one of those and view it at whatever size you like. It passes its content through, so you can chain it. ### Compare An A versus B image comparison with a draggable wipe divider. Wire two images in and drag the slider to compare them. No output. ## Organizational nodes These help you annotate and structure the canvas. They have no handles. ### Comment A draggable sticky note for annotating the canvas. It can grow a pointer arrow that pins to a spot or anchors to a specific node and follows it. ### Note A simple sticky text note. ### Section The labeled frame that contains other nodes. See the sections part of [The Canvas Editor](https://docs.3daistudio.com/flow/interface). ## Custom nodes You can create your own reusable nodes: - **Pre Prompt** - a fixed text prefix or suffix. - **Custom Image Edit** - a locked model, style prompt, and default settings baked into one node, so you just wire an image and run. - **Group** - a captured sub graph (several nodes and their wires) that you can stamp onto any canvas as a unit. Custom nodes appear in the Add menu under their own section. Saving a selection as a group is one click from the selection toolbar. ## Quick handle reference A condensed table of what each main node takes and produces: | Node | Inputs | Outputs | | --- | --- | --- | | Prompt | none | text | | Combine Prompts | up to 10 text | text | | LLM | prompt, images | text | | Generate Image | prompt | image | | Edit Image / Style Reference | images or image set, prompt | image | | Remove Background | image | image | | Pose Changer / Optimize Image | one image | image | | Image to 3D / Multi View to 3D / Text to 3D | image / views / text | mesh | | Texture / Remesh / Convert | mesh (+ optional prompt or image) | mesh | | Image | (source) | image | | Images | none | per slot images | | Moodboard | images | image set | | Multi View | image | front, left, right, back | | Render Views | mesh | image set and per view images | | Combine to Grid | images or image set | image, layout | | Split Grid | image, layout | image set and per cell images | | Split into Parts | image | per part images, image set | | Collect | one image or mesh | All results image set, out1...outN | | Batch | images, prompts, or meshes | the same type, run once per item | | Switch | several of one type | that type | | Seed | none | number | | Variable | none | text or number | | Viewer | image, mesh, or text | passthrough | | Compare | two images | none | | 3D Model | (source) | mesh | | File | result | none | ## Frequently Asked Questions ### What do the colors on the wires mean? Each color is a data type. Red is text, blue is a single image, indigo is a group of images, green is a 3D mesh, amber is a number, violet is language model text, and cyan is a viewer passthrough. You can only connect two handles of the same color, which keeps pipelines valid. ### Why does my generation node not have the input I expected? Generation node inputs are derived from the model you select. Different models accept different inputs (some take an image set, some take individual image slots, some take only text). Switch the model in the node's picker and the input handles update to match. ### What is the difference between Multi View and Render Views? Multi View takes a single image and uses AI to imagine four orthographic views of it (this costs credits). Render Views does the opposite: it takes an existing 3D model and renders turnaround images of it locally in your browser with no credits. Multi View feeds into 3D generation; Render Views extracts images from a finished model. ### What is a Switch node for? A Switch routes one of several inputs to its output, acting as a fork in your pipeline. In manual mode you pick the active branch; in auto mode the first input with content wins. When you run the flow, only the selected branch executes, so you can keep several alternatives wired up and choose between them without deleting anything. ### How do I reuse a part of my pipeline in other flows? Select the nodes you want to reuse, then click Save group in the selection toolbar. This captures them (and their wiring) as a custom group node. It then appears in your Add menu and can be stamped onto any canvas as a single unit. ### What does bypassing a node do? Bypassing (muting) a generation node makes it pass its input straight through to its output without running. The pipeline behaves as if the node were not there. It is the quick way to temporarily skip a step without unwiring it. Select the node and press B to toggle. [Open Flow](https://www.3daistudio.com/Flow) - [Try Image to 3D](https://www.3daistudio.com/ImageTo3D) - [3D AI Studio home](https://www.3daistudio.com) --- # Running Flows Source: https://docs.3daistudio.com/flow/running-flows Running is where your pipeline actually does work. This page explains the different ways to run, how a run plays out, how the run queue works, and how to turn a flow into a simple reusable app. ## The three ways to run There are three buttons that trigger work, each with a different scope. ### 1. Run a single node Every generation node has a **Run node** bar at the bottom. Clicking it runs just that one node using its current inputs. It does not touch upstream nodes. Use it when you only want to regenerate one step, for example to roll a new variation of an image you already have inputs for. ### 2. Run flow to here Every generation node also has a **Run flow to here** action in its header. This runs everything that node depends on, in the correct order, ending with that node. Use it to build up to a specific point in your pipeline without running steps further downstream. ### 3. Run the whole flow The big **Run** button in the top right runs the whole pipeline up to its outputs. Clicking it opens the **run panel**, where you confirm inputs and start the run. > [!NOTE] > There is always one launch surface: the run panel. A node never quietly runs a multi step pipeline on its own. When you press Run or "Run flow to here", the run panel opens, shows you the plan, and you start it from there. This keeps runs predictable. ## What happens when you run When a run starts, Flow does the following automatically: 1. **Works out the order.** It follows your wires and figures out the correct order automatically, so that everything a node needs is produced before that node runs. If your wiring accidentally forms a loop, Flow refuses to run and tells you there is a loop. 2. **Sees through pass through nodes.** Bypassed nodes, viewer nodes, and switches do not run themselves; Flow looks through them to find the real upstream producer. For a manual switch, only the selected branch is followed. 3. **Runs one node at a time, in order.** Each node gathers its inputs from the connected upstream nodes, runs that step, and waits for it to finish before starting the next one. Running is deliberately sequential so it is predictable, reliable, and easy to follow. 4. **Feeds results forward.** Each finished result is stored on the node and becomes the input for the next step. You can watch all of this happen: nodes light up with a status ring, a number badge shows their position in the order, and wires animate as data flows along them. The Flow Monitor panel in the bottom left shows a live list of steps with their status, the elapsed time, and a Stop button. ![A node in its generating state during a flow run](https://docs.3daistudio.com/assets/images/docs/flow/flow-node-generating.webp) *A node mid run. While it generates, the next steps wait their turn and the wires animate to show data flowing.* ### Run modes: all versus missing When you start a run you can choose: - **All** - run every node in the plan, even ones that already have a result. - **Missing** - run only nodes that do not have a result yet, plus anything downstream of them (since changing an upstream step makes downstream results stale). "Missing" is the efficient choice when you have already generated part of a pipeline and only want to fill in the rest. ## The run panel The run panel is the control center for running. It opens when you press Run or "Run flow to here". It shows: - A **plan preview**: which steps will run and in what order, and a warning if any steps are missing inputs and will be skipped. - An **inputs form** built from your promoted inputs (see below). - A **credit estimate** for the run. - The **queue and activity** of past and pending runs. Press Run inside the panel to queue the run. ![The Flow run panel showing input fields, the plan, and a credit estimate](https://docs.3daistudio.com/assets/images/docs/flow/flow-run-modal.webp) *The run panel. Promoted inputs become a simple form, and you see the plan and credit estimate before starting.* ## The run queue Flow runs are queued, similar to a print queue. Each queued item captures a snapshot of the inputs you set, so you can change the inputs and queue another run without waiting for the first to finish. - Runs execute **one at a time**, in order. - You can **queue multiple** runs. - Each item shows its status: queued, running, done, partial, failed, or cancelled. - You can **cancel** a queued run, **stop** the active one, and **clear** finished items. This makes it easy to fire off several variations (for example three different prompts) and let them process while you do other things. ## Batch runs A **Batch** node lets one step run **once per item**, so you can generate or transform a whole set in a single go. Wire a Batch (a collection of images, prompts, or 3D models) into a generation or derivation node and that node's Run button becomes **Run N×**, where N is the number of items. See the [Batch node](https://docs.3daistudio.com/flow/nodes#batch) for how to fill one. - Items run with a sensible **concurrency** per type (for example a couple of 3D generations at a time, more for fast image edits), starting the next as each one finishes. - Failed items are **not** retried automatically, and one failure does not stop the rest. - Progress shows live as **k / N** on the node and in the run monitor, and every result lands in the node's history filmstrip. - To pass *all* the results to a following step, add a **Collect Results** after the batched node. Chaining two batched stages looks like: Batch to Style Reference to Collect to Batch to Image to 3D. See [Collect Results](https://docs.3daistudio.com/flow/nodes#collect-results). ### Batch and credits A batched run costs credits **per item**, so N items is N generations. The run panel shows the true total before you run, for example "~18 credits per run, 6 generations", and a banner notes when a run fans out into a batch. For several *variations of one prompt*, use the model's image count setting instead of a Batch. ## Turning a flow into an app: promoted inputs and outputs This is one of the most powerful features in Flow. You can mark certain nodes as the **inputs** and **outputs** of your pipeline, which turns the run panel into a simple form. Anyone can then run your flow by filling in the form, without touching the node graph at all. ### Promoted inputs Promote a node to make it appear as a field in the run panel. These node types can be promoted: - **Prompt** nodes become text fields. - **Image** nodes become image upload fields. - **Variable** nodes become text, number, or choice fields. - **Switch** nodes become a choice between branches. Each promoted input gets a label and an order so the form reads logically. ### Promoted outputs Mark a generation node as a **promoted output** to make it a target of the run. When you run the flow, Flow runs everything needed to produce all the marked outputs. If you do not mark any outputs, Flow falls back to running the natural end points of your pipeline, so an unconfigured flow still just runs. > [!NOTE] > A flow with promoted inputs shows the green **Run** badge on its dashboard card and can open directly into run mode. This is how you build a reusable tool: set it up once with the right inputs and outputs, then run it (or share it) as a simple form, no wiring required. You can promote inputs and outputs from a node's header actions, or ask the AI assistant to do it for you. ## Reproducibility with seeds To get consistent or repeatable results, add a **Seed** node and wire it into your generators. Set it to "fixed" to reuse the same seed every run (so the same prompt gives the same image), or "new each run" to get a fresh variation each time. When a run starts, Flow rolls the seed once and every generator sharing it uses the same value. ## Results, history, and outputs - Each generation node keeps its result **on the node**, with a **history filmstrip** of previous results. Click any thumbnail to make it the active result. - Generating more than one image at a time sends the extra results into the filmstrip as variations. - The **Outputs** drawer (top right) lists everything the flow has ever generated, with the node and run that produced each one. From there you can locate the producing node, set a result as active, add it to the canvas as a new node, or download it. ## Errors and retries Flow is built to keep a run going even when one step fails: - If a node fails, the steps that depend on it are marked **skipped** rather than crashing the whole run. - A failed node shows an error with a **Retry** button. Retrying starts a fresh attempt. - If you reload the page mid run, finished results are still there, and any steps still in progress carry on where they left off. - At the end of a run you get a summary, for example "Flow complete, 3 generated, 1 failed, 1 skipped". ## Frequently Asked Questions ### What is the difference between running a node and running the flow? Running a single node (the Run node bar at the bottom of a node) executes just that one step with its current inputs. Running the flow (the Run button top right) executes the whole pipeline in dependency order, ending at your outputs. There is also "Run flow to here" in a node's header, which runs everything up to and including that node. ### How does Flow know what order to run nodes in? It follows your wires and works out which steps depend on which, then runs them in the right order so everything a node needs is ready before that node runs. This is automatic. If your wiring accidentally forms a loop, Flow refuses to run rather than getting stuck. ### Can I queue several runs at once? Yes. Flow has a run queue, so you can line up several runs. Each queued run snapshots the inputs you set, so you can change a prompt or image and queue another run immediately. Runs execute one at a time, and you can cancel queued ones or stop the active one. ### How do I run the same step on many inputs at once? Use a **Batch** node. Put your images, prompts, or 3D models into it (upload them or wire sources in), then connect it to the generation or derivation node, which then runs once per item. Add a Collect Results after it to gather all the outputs. See [Batch runs](https://docs.3daistudio.com/flow/running-flows#batch-runs). For several variations of a single prompt, use the model's image count setting instead. ### What does "missing" run mode do? Missing mode runs only the nodes that do not have a result yet, plus anything downstream of them. It is the efficient choice when part of your pipeline is already generated and you only want to complete the rest. All mode reruns everything regardless of existing results. ### How do I make a flow that other people can run without editing nodes? Promote the inputs and outputs. Mark your Prompt, Image, Variable, or Switch nodes as promoted inputs (they become form fields in the run panel) and mark your final generation nodes as promoted outputs. The flow can then be run as a simple form. Flows set up this way show a green Run badge on their dashboard card. ### Why was one of my steps skipped during a run? A step is skipped when something it depends on failed, or when it is missing a required input. Flow cascades a skip downstream so a single failure does not crash the whole run. Check the failed node's error message, fix the input or retry that node, then run again in missing mode to fill in the rest. ### Do bypassed and switch nodes run? No. Bypassed nodes and viewer nodes do not execute; Flow looks through them to the real upstream producer. A switch does not run either, it just routes one branch through. In manual mode only the selected branch of a switch is executed during a run. [Open Flow](https://www.3daistudio.com/Flow) - [Try Image to 3D](https://www.3daistudio.com/ImageTo3D) - [3D AI Studio home](https://www.3daistudio.com) --- # AI Assistant (Bob) Source: https://docs.3daistudio.com/flow/ai-assistant Bob is the AI assistant built into Flow. Instead of placing and wiring nodes yourself, you can describe what you want in plain English and Bob builds or edits the pipeline for you. Bob can create whole workflows, change existing nodes, rewire connections, and answer questions about how things work. ![The Bob AI assistant building a pipeline on the canvas from a plain English description](https://docs.3daistudio.com/assets/images/docs/flow/flow-ai-builder.webp) *Describe what you want and Bob proposes a complete pipeline, which you can apply or discard.* ## Opening Bob Click the assistant button in the top right cluster of the editor. A chat panel opens that you can dock to an edge or float anywhere on screen. Opening the chat does not cost anything; you spend credits only on the messages you send. You can also reach Bob through the **Build with AI** option in the new flow launcher. In that case Bob opens automatically and starts building from your description as soon as the editor loads. ## What Bob can do Bob works in two complementary ways, often in the same response. ### Build new pipelines Describe a workflow and Bob proposes a complete set of new nodes, already wired together, laid out neatly, and grouped in a section. For example: > "Generate a fantasy sword concept image, create four turnaround views of it, then convert it into a 3D model." Bob will propose a Prompt node, a Generate Image node, a Multi View node, and a Multi View to 3D node, all connected in the right order. ### Edit what is already there Bob can also change your existing canvas. It can: - Switch a node's model. - Change a node's settings. - Connect or disconnect nodes. - Insert a new node in the middle of an existing wire. - Rename, group into a section, bypass, or delete nodes. - Promote inputs and mark outputs (turning your flow into an app). - Drop an image you attach to the chat into an Image node. ### Answer questions and guide you You can just ask Bob how something works, or what node to use for a goal. Bob can also highlight a specific node on your canvas to point you to it, and can open the feedback form for you if you hit a problem. ## How proposals work (review before apply) Bob never changes your canvas without showing you first. Every response that includes changes comes as a **proposal card** with a clear choice: - **Apply / Keep** to accept the changes. - **Discard** to reject them. The whole proposal applies as a **single undo step**, so if you change your mind you can undo it all at once with Ctrl+Z. This is a core safety principle: Bob can only do things you could do by hand, and nothing happens until you confirm. ### Watch Bob build When Bob is building a new pipeline, you can literally watch the nodes appear on the canvas as ghosts while it works. When it finishes, the nodes settle and you confirm with Keep or Discard. ## Apply modes: Ask versus Auto You can choose how proposals are applied, set per flow: - **Ask** (the default): Bob shows you each proposal and you click to apply it. - **Auto**: proposals are applied automatically right after the preview. It is still one undo step, so you can always reverse it. There is also an **auto arrange** setting that controls whether Bob tidies the layout of the parts of the canvas it touches. ## Reasoning effort and credits Each **message** you send costs credits, priced by the reasoning effort you choose: | Effort | Good for | Cost per message | | --- | --- | --- | | **Fast** | Small edits and simple builds | 1 credit | | **Smart** | Most builds | 1 credit | | **Genius** | Large, complex pipelines | 2 credits | Higher effort gives Bob more time to think, so it can plan and build larger, more complex pipelines in one go. Your choice is remembered between sessions. Each chat also has a message cap; when you reach it, start a new chat. ## How Bob understands your canvas Bob can see the nodes on your canvas and how they are connected, so its suggestions fit what you are actually building rather than being generic. You can also **attach images** to a message for Bob to look at and place on the canvas. ## Self correction If a proposal does not apply cleanly (for example a connection that no longer fits after other edits), Bob notices and fixes it on its next try, so you rarely have to explain what went wrong. ## Tips for working with Bob - **Be specific about the goal, not the nodes.** "Make a pipeline that turns a photo into a 3D printable model" works better than trying to name every node. - **Build in steps for big pipelines.** Ask for the core chain first, then ask Bob to add refinements. - **Use Ask mode while learning** so you can see exactly what each proposal does before applying it. - **Combine with manual editing.** Let Bob lay down the structure, then fine tune settings yourself. - **Ask questions.** Bob is also a guide, not just a builder. ## Frequently Asked Questions ### What is Bob in Flow? Bob is the AI assistant inside the Flow node editor. You describe what you want in plain English and Bob builds the pipeline, edits existing nodes, rewires connections, or answers questions. Every change is shown as a proposal you can apply or discard, and applies as a single undo step. ### Does Bob change my canvas automatically? Only if you turn on Auto apply mode. By default Bob is in Ask mode, where it shows you a proposal card and you click Apply or Discard. Either way, the whole change is one undo step, so you can always reverse it with Ctrl+Z. ### How much does the AI assistant cost? Opening the chat does not cost anything. Each message costs credits based on the reasoning effort you pick: Fast and Smart are 1 credit, Genius is 2 credits. Higher effort lets Bob build larger pipelines in one response. ### Can Bob build a whole pipeline from one sentence? Yes. Describe the workflow you want, for example "generate a character image, make four views, and turn it into a 3D model", and Bob proposes all the nodes wired together. For very large pipelines, use the Genius effort or build in a few steps. ### Can I attach an image to the assistant? Yes. You can attach images to a message. Bob can see them (so it can reason about them) and can place them into an Image node on your canvas as part of a proposal. ### What is the difference between Ask and Auto apply mode? In Ask mode (default), Bob shows each proposal and waits for you to click Apply or Discard. In Auto mode, proposals are applied automatically right after the preview, still as a single reversible undo step. Use Ask while you are learning so you can review each change first. [Open Flow](https://www.3daistudio.com/Flow) - [Try Image to 3D](https://www.3daistudio.com/ImageTo3D) - [3D AI Studio home](https://www.3daistudio.com) --- # Templates & Sharing Source: https://docs.3daistudio.com/flow/templates Once you have built a pipeline you like, you do not have to rebuild it. Flow lets you save it as a template, share it with a link, publish it to the public gallery, and keep a history of versions you can roll back to. This page covers the whole lifecycle. ## The vocabulary Four similar sounding words mean different things in Flow: | Term | What it is | | --- | --- | | **Flow (canvas)** | Your editable working document, the pipeline you are building. | | **Template** | A saved snapshot of a flow (its nodes plus a title, description, category, and cover images) that can be reused, shared, and published. | | **Version** | A saved point in your flow's history that you can restore. Different from a template. | | **Custom node / group** | A small reusable piece (one node or a captured sub graph), not a whole template. | ## Saving a flow as a template Click **Save as template** in the bottom left of the editor. A dialog opens with a live preview of the template card on the left and a form on the right. Fill in: - **Title** and **description** (both required). - An **icon**. - A **category** and **difficulty** (both required). - **Tags** (up to twelve) to help people find it. - **Cover images** - pick up to four of your flow's output images to show on the card. Click **Save private template**. The result is a **private draft** that only you can see, listed under My Templates on the dashboard. ![The Save as template dialog with a live preview card and a metadata form](https://docs.3daistudio.com/assets/images/docs/flow/flow-save-template.webp) *The Save as template dialog. The live preview on the left updates as you fill in the form on the right.* ## Private drafts versus public copies A template you own can be in two states: - A **private draft** is visible only to you. You can keep editing it and submit it to the gallery later. - A **public copy** is a detached snapshot you submitted for the gallery. Your private draft stays separate, so submitting does not change your working draft. Each public copy carries a status: **In review**, **Live** (published), or **Not approved** (with a reason). ## Using a template Open a template's detail view (click its card) to see a large preview of the pipeline and its information. The main action is **Use this template**, which clones the template into a brand new flow of your own. The original is never changed. ![A template detail view with a large pipeline preview and the Use this template button](https://docs.3daistudio.com/assets/images/docs/flow/flow-template-preview.webp) *The template detail view, with a preview of the pipeline and the actions to use, share, or submit it.* If the template was set up as an app (it has promoted inputs), using it can open directly into run mode so you can fill in the form and run it right away. From the detail view you can also **Favorite** a template (heart) and, for public templates, **Report** it if something is wrong. ## Sharing a flow with a link Click **Share link** (in the editor's bottom left bar, or from a flow card on the dashboard). This mints a short lived link to your flow. - Choose a duration: **10 minutes**, **1 hour**, or **24 hours**. - Choose **Allow copying** (recipients can clone it into their own account) or leave it off for **view only**. - Only one active link exists per item; creating a new one replaces the old. - You can **Revoke** a link at any time. Anyone with the link opens a clean, full screen, read only view of your flow. If copying is allowed, they get a **Use this flow** button to clone it. > [!NOTE] > Share links expire automatically, so you stay in control of who can open your flow. ## Submitting to the public gallery To share a template with the whole community: 1. Save it as a private template. 2. Open its detail view and click **Submit to gallery**. This creates a public copy with the status **In review**. It is then reviewed before going live: - **Approved** templates go **Live** and appear in the Public Gallery, where anyone can use them. - **Rejected** templates show as **Not approved** with a reason, so you can fix and resubmit. If you spot a problem with a published template, you can report it from its detail view. ## Browsing the gallery The **Public Gallery** tab on the dashboard is the community library. You can search by name, filter by category, and sort by New, Trending, Most Used, Official, or Staff Picks. Hover a card to preview its description, favorite it with the heart, or open it to use it. ## Version history Version history is separate from templates. It is the long term companion to undo and redo, working across sessions. Open **Version history** from the bottom left bar. You get: - **Manual versions** you save with a name, marking a milestone. - **Automatic checkpoints** saved periodically as you work. For any version you can: - **Restore** it in place. Your current state is saved first, so a restore is always reversible. - **Open as a copy** (branch it into a new flow). - **Rename**, **Pin** (so it is never auto deleted), or **Delete** it. > [!NOTE] > Use version history to experiment with confidence: save a version before a big change, and restore it if the change does not work out. ## The full lifecycle at a glance 1. **Create** a flow (blank, with AI, or from a template). 2. **Build** it, saving named versions along the way. 3. **Save as template** to make a private, reusable snapshot. 4. **Use** the template anytime to spin up fresh copies. 5. **Share** it privately with a short lived link. 6. **Submit** it to the gallery to share with everyone. 7. A reviewer **approves** it and it goes Live for the community. ## Frequently Asked Questions ### What is the difference between a template and a version? A template is a reusable snapshot of a flow with its own title, description, and cover images, meant to be cloned into new flows or published. A version is a saved point in a single flow's history that you can restore. Templates are for reuse and sharing; versions are for going back in time within one flow. ### How do I reuse a flow I built? Save it as a template (Save as template in the bottom left of the editor). It becomes a private draft under My Templates. Open its detail view and click Use this template to clone it into a new flow whenever you want, without changing the original. ### How do I share a flow with someone? Click Share link in the editor or on a flow card. Choose how long the link lasts (10 minutes, 1 hour, or 24 hours) and whether recipients can copy it or only view it. Send them the link. They open a read only view, and if copying is allowed they can clone it into their own account. ### How do I publish a template to the public gallery? First save it as a private template. Then open its detail view and click Submit to gallery. It goes into review with the status In review. Once a reviewer approves it, it goes Live and appears in the Public Gallery for everyone to use. ### Can I get back an earlier version of my flow? Yes. Open Version history from the bottom left bar. You can restore any manual version you saved or any automatic checkpoint. Restoring saves your current state first, so it is always reversible. You can also open a version as a separate copy instead of overwriting your current flow. [Open Flow](https://www.3daistudio.com/Flow) - [Try Image to 3D](https://www.3daistudio.com/ImageTo3D) - [3D AI Studio home](https://www.3daistudio.com) --- # Models, Credits & Downloads Source: https://docs.3daistudio.com/flow/models-credits-downloads This page covers the practical side of generation in Flow: choosing models, adjusting their settings, what things cost, and how to view and download your results. ## Choosing a model Generation nodes (Generate Image, Edit Image, the 3D nodes, and so on) let you pick which AI model does the work. Flow offers a curated set of the platform's best models, organized so you can find the right one quickly. - **Image nodes** use a small, high quality set of image models. - **3D nodes** use a categorized picker covering image to 3D, multi view to 3D, and text to 3D, with filters and search. When you pick a model, the node's input handles automatically update to match what that model accepts. For example, switching to a model that takes multiple views will change the inputs to front, left, right, and back slots. This means you never have to guess which inputs a model needs. > [!NOTE] > The same models you know from the standalone tools (such as Prism, Hunyuan, and the image models) are available in Flow. They behave the same way; Flow simply wires them into a pipeline. You can preview model quality and costs on the standalone [Image to 3D page](https://www.3daistudio.com/ImageTo3D) before building them into a flow. ## Adjusting parameters Each model exposes its own settings in the node, so every model shows exactly the controls it supports: - **Toggles** for on/off options (like PBR materials). - **Dropdowns or pill selectors** for choices (like quality or aspect ratio). - **Sliders** for numeric values (like face count), with safe minimum and maximum limits. - **Advanced settings** are tucked behind a "Show advanced" toggle so the node stays simple by default. If you change a value away from its default, a small reset control appears so you can put it back. ## Parameter binding (wiring a setting) Sometimes you want a setting to be driven by another node rather than set by hand. Flow lets you **bind** a parameter, which turns it into an input handle on the node. Then you can wire a value into it. The classic example is a **Seed** node wired into a model's seed parameter, so several generators share one seed for consistent results. You can also drive a setting from a **Variable** node, which is useful when you promote that variable into the run panel as a form field. Numeric settings accept a number wire; text and choice settings accept a text wire. ## Credits You spend credits only when you run nodes that call AI models. Building, wiring, and arranging the canvas costs nothing. - Each model shows its **credit cost** in the picker and in the Node Library. - The **run panel** shows a credit estimate for a whole run before you start it. - Generating multiple images at once multiplies the cost. A multi view step counts as four image generations. - A **Batch** multiplies cost per item: a step fed by a batch of 6 runs six times, so it costs six generations. The run panel reflects this and shows the generation count (for example "6 generations") before you run. See [Batch runs](https://docs.3daistudio.com/flow/running-flows#batch-runs). - Local steps like **Render Views** (rendering turnaround images from a 3D model in your browser) and pure logic nodes (Switch, Seed, Variable, Combine Prompts, Batch) cost nothing on their own. Your current balance is always shown in the top right of the editor. ## Downloads Every download in Flow goes through one place, so you get a consistent experience: a small downloads tray appears in the corner showing progress for each file, and you can cancel or dismiss items. ## The Outputs shelf Open **Outputs** from the top right cluster to see a gallery of everything this flow has ever generated, newest first. Each tile is tagged with the node and run that produced it. Hover a tile for actions: ![The Outputs shelf listing every result the flow has generated](https://docs.3daistudio.com/assets/images/docs/flow/flow-outputs-panel.webp) *The Outputs shelf collects every result the flow has produced, with quick actions on each tile.* - **Download** - save the result to your computer. - **Locate node** - jump the camera to the node that made it. - **Set as active result** - make this the node's current output. - **Add to canvas** - drop the result onto the canvas as a new Image or 3D Model node. You can filter by type (all, images, 3D) and by a specific run. In progress generations show a "Generating" state and update live. ## The Content Viewer The **Content Viewer** is a panel that mirrors whatever node you currently have selected, shown larger. Select an image node and it shows the image; select a 3D node and it shows a rotatable model; select a text node and it shows the text. It is the quick way to inspect a result closely without zooming the canvas. It also has a history strip so you can flip through a node's previous results. ## Preview quality (fast versus sharp) To keep large canvases smooth, image previews can be shown as lightweight thumbnails. The **resolution toggle** in the toolbar switches between: - **Fast** - lightweight thumbnails, best for big pipelines. - **Sharp** - full resolution previews. This affects display only. Wires, downloads, and large viewers always use the full quality asset regardless of the toggle, so your actual outputs are never reduced in quality. ## Frequently Asked Questions ### Which models can I use in Flow? Flow offers a curated selection of 3D AI Studio's best image and 3D models, including the Prism and Hunyuan 3D families and the top image generation and editing models. The 3D picker is organized by category (image to 3D, multi view to 3D, text to 3D) with filters and search. Picking a model automatically updates the node's inputs to match. ### How much do generations cost in Flow? The same as the standalone tools. Each model shows its credit cost, and the run panel estimates the total cost of a run before you start. You only spend credits when you run model based nodes. Generating multiple images multiplies the cost, and a multi view step counts as four generations. Logic nodes and local rendering cost nothing. ### How do I download a result? Use the download button on a node's result, on a tile in the Outputs shelf, or in the Content Viewer. A downloads tray shows the progress of each file. ### What is parameter binding? Parameter binding turns a model setting into an input handle so you can wire a value into it instead of setting it by hand. The most common use is wiring a Seed node into a model's seed for consistent results, or driving a setting from a Variable node that is promoted into the run panel. Numeric settings take a number wire; text and choice settings take a text wire. ### Does the fast preview setting reduce my output quality? No. The fast versus sharp toggle only changes how previews are displayed on the canvas to keep large flows smooth. The actual data passed along wires, used in large viewers, and downloaded is always full quality. Switch to sharp anytime you want to inspect previews at full resolution. ### Where can I see everything a flow has generated? Open the Outputs shelf from the top right of the editor. It lists every result the flow has produced, newest first, tagged by node and run. From there you can download a result, jump to the node that made it, set it as the active result, or add it to the canvas as a new node. [Open Flow](https://www.3daistudio.com/Flow) - [Try Image to 3D](https://www.3daistudio.com/ImageTo3D) - [3D AI Studio home](https://www.3daistudio.com) --- # MCP Connector Overview Source: https://docs.3daistudio.com/mcp/overview The 3D AI Studio MCP connector is a [3D model generation MCP server](https://www.3daistudio.com/MCP): it lets your AI assistant create 3D models and images for you. You connect Claude, ChatGPT, Cursor, Claude Code, Codex or another assistant to your 3D AI Studio account once, and from then on you can simply ask in the chat: > "Turn this sketch into a 3D model and export it as an STL for printing." The assistant picks the right tool, starts the generation on 3D AI Studio, waits for it to finish, and shows you the result. Everything it creates is saved to your normal [3D AI Studio dashboard](https://www.3daistudio.com/Dashboard), exactly as if you had made it on the website. > [!NOTE] > **Connector URL:** `https://mcp.3daistudio.com/mcp` > > This is the only address you need. It is the same for every user. Your account is connected through a normal sign in screen, so there is no API key to create or copy. ![Claude with the 3D AI Studio connector: the user asks for an orc warrior and the finished 3D model appears in an interactive viewer in the chat](https://docs.3daistudio.com/assets/images/docs/mcp/mcp-chat-viewer.webp) *Ask for a model in the chat, and a real, textured 3D model comes back. In Claude it opens in an interactive viewer, ready to download or open in 3D AI Studio.* ## What is MCP? MCP stands for **Model Context Protocol**. It is an open standard that lets AI assistants use external tools. A tool provider (in this case 3D AI Studio) publishes a list of things it can do, and the assistant decides when to use them while it talks with you. On its own, an assistant like ChatGPT or Claude cannot generate real 3D models. At best it can write code for very simple shapes. With the 3D AI Studio connector added, it can use the same AI models you use on the website and hand you a proper textured mesh with downloadable files. ## What you can do Once connected, you can ask your assistant to: - **Generate 3D models** from a text description, a single image, or several views of the same object (Prism 3.1, P2, P1, Hunyuan 3.1, Meshy 7.1). - **Generate and edit images** (Nano Banana 2.1, GPT Image 2, and more), for example to design a character first and then turn it into 3D. - **Make a low poly version** of a model for games, AR and VR, or **unwrap its UVs**. - **Convert a model** to GLB, OBJ, FBX, STL, PLY, USDZ or 3MF, optionally scaled to a target size for 3D printing. - **Create PBR materials** from a text description or a photo (base color, normal, roughness, metalness and height maps). - **Remove image backgrounds.** - **Browse your library** and **organize results into projects** in your dashboard. - **Check your credit balance** and compare model costs before generating. The full list, with costs and example prompts, is on [What You Can Do](https://docs.3daistudio.com/mcp/capabilities). ## Requirements | You need | Details | | --- | --- | | A 3D AI Studio account on a **paid plan** | The connector is part of every paid plan. Free accounts can see the setup but are asked to upgrade when connecting. [See plans](https://www.3daistudio.com/Pricing) | | An AI assistant that supports **remote MCP servers** | Claude, ChatGPT, Claude Code, Cursor, Codex, VS Code with GitHub Copilot, OpenCode and most other MCP clients. Some assistants only offer custom connectors on their own paid plans | | A web browser | Only for the one time sign in when you connect | Generations use your **normal 3D AI Studio credits**, at the same prices as on the website. There is no separate subscription or wallet for the connector. ## Supported assistants | Assistant | Where it runs | Setup | Result preview | | --- | --- | --- | --- | | [Claude](https://docs.3daistudio.com/mcp/setup/claude) | Web and desktop | Add a custom connector in Settings | Interactive 3D viewer in the chat | | [ChatGPT](https://docs.3daistudio.com/mcp/setup/chatgpt) | Web and desktop | Add a connector in Settings | Result card or link | | [Claude Code](https://docs.3daistudio.com/mcp/setup/claude-code) | Terminal | One command | Link to the result | | [Cursor](https://docs.3daistudio.com/mcp/setup/cursor) | Editor | One config entry | Link to the result | | [Codex](https://docs.3daistudio.com/mcp/setup/codex) | CLI, IDE extension and app | One command | Link to the result | | [VS Code (Copilot)](https://docs.3daistudio.com/mcp/setup/vs-code) | Editor | One config entry | Link to the result | | [OpenCode](https://docs.3daistudio.com/mcp/setup/opencode) | Terminal | One config entry | Link to the result | | [Other MCP clients](https://docs.3daistudio.com/mcp/setup/other-clients) | Anywhere | Add a remote server with the URL | Depends on the client | AI apps update their MCP settings often, so if a guide's menus or commands do not match what you see, ask your assistant how it adds a remote MCP server and point it to [our agent instructions](https://docs.3daistudio.com/mcp/agent-setup). Coding agents can usually do the whole setup for you. How results are shown depends on the assistant. Clients that support interactive cards (like Claude) show finished models in a 3D viewer you can spin right in the chat. Terminal and editor clients show a link to the result instead. Either way, the model is always in your dashboard. ## How it works 1. **Connect once.** Add the connector URL to your assistant, sign in with your 3D AI Studio account, and approve access. See the [Quick Start](https://docs.3daistudio.com/mcp/quick-start). 2. **Ask in plain language.** For example: "make a low poly fox for my game" or "turn this product photo into a 3D model". 3. **The assistant does the work.** It chooses a model, starts the generation, and checks on it until it is done. 3D models usually take 1 to 5 minutes. 4. **Get your result.** The model appears in the chat (or as a link) and in your dashboard, where every editing tool, export format and project feature works as usual. > [!NOTE] > **Prefer to let your AI do the setup?** Developer tools like Claude Code, Codex and Cursor can configure the connector themselves. Send your agent the [agent setup page](https://docs.3daistudio.com/mcp/agent-setup) and it will take care of the rest. ## Where to go next - **[Quick Start](https://docs.3daistudio.com/mcp/quick-start)** - connect your first assistant in about two minutes. - **[Set Up With Your Agent](https://docs.3daistudio.com/mcp/agent-setup)** - a copy and paste brief for coding agents. - **[Setup Guides](https://docs.3daistudio.com/mcp/setup/claude)** - step by step instructions for each assistant. - **[What You Can Do](https://docs.3daistudio.com/mcp/capabilities)** - every tool, model and cost, with example prompts. - **[Images & Uploads](https://docs.3daistudio.com/mcp/images-and-uploads)** - how to give your assistant a photo or sketch. - **[Credits & Limits](https://docs.3daistudio.com/mcp/credits-and-limits)** - how spending, confirmations and limits work. - **[Managing Connections](https://docs.3daistudio.com/mcp/managing-connections)** - see, fix and disconnect your assistants. - **[Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting)** - fixes for every common message. - **[Security & Privacy](https://docs.3daistudio.com/mcp/security-privacy)** - what an assistant can and cannot do with your account. ## Frequently Asked Questions ### Can ChatGPT and Claude generate 3D models? Not on their own, but they can with the connector added. See how to [generate 3D models in ChatGPT and Claude](https://www.3daistudio.com/MCP) for examples, or jump straight to the [ChatGPT](https://docs.3daistudio.com/mcp/setup/chatgpt) or [Claude](https://docs.3daistudio.com/mcp/setup/claude) setup guide. ### Do I need an API key? No. You sign in through a normal 3D AI Studio login screen and approve access. No key is created, copied or stored in a config file. ### Does the connector cost extra? No. It is included in every paid plan. Generations cost the same credits as on the website. ### Where do the generated models go? Into your normal dashboard, tagged so you can see which ones came from an assistant. You can download, edit, convert and organize them like any other model. ### Can I disconnect an assistant later? Yes, at any time in [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp). It stops working immediately. See [Managing Connections](https://docs.3daistudio.com/mcp/managing-connections). [Open AI Assistants settings](https://www.3daistudio.com/Dashboard?settings=mcp) - [MCP on 3daistudio.com](https://www.3daistudio.com/MCP) - [Pricing](https://www.3daistudio.com/Pricing) --- # MCP Quick Start Source: https://docs.3daistudio.com/mcp/quick-start This page gets you from zero to your first AI generated 3D model inside your assistant in about two minutes. If you want the exact clicks for one specific assistant, jump to its [setup guide](https://docs.3daistudio.com/mcp/quick-start#step-by-step-guides-for-each-assistant). New to 3D AI Studio? See the [3D model generation MCP](https://www.3daistudio.com/MCP) for what it can do and example results. ## What you need - A **3D AI Studio account on a paid plan** ([see plans](https://www.3daistudio.com/Pricing)). - An **AI assistant that supports remote MCP servers**, such as Claude, ChatGPT, Claude Code, Cursor, Codex, VS Code with Copilot, or OpenCode. These two values are all you will ever type during setup: | Setting | Value | | --- | --- | | Connector URL | `https://mcp.3daistudio.com/mcp` | | Server name (for tools that ask for one) | `3daistudio` | > [!WARNING] > Use the URL exactly as shown, including `mcp.` at the start and `/mcp` at the end. `www.3daistudio.com` and older guides that end in `/sse` will not work. ## Connect in four steps ### Step 1: Check your plan The connector is included in every paid 3D AI Studio plan. You can confirm your plan, and see the connector URL, in [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp). ### Step 2: Add the connector to your assistant Add a new remote MCP server (some apps call it a "custom connector") with the connector URL. How you do that depends on the app: | Assistant | Where to add it | | --- | --- | | Claude (web and desktop) | Settings → Connectors → Add custom connector, then paste the URL. [Full guide](https://docs.3daistudio.com/mcp/setup/claude) | | ChatGPT | Settings → Connectors → Create, then paste the URL. [Full guide](https://docs.3daistudio.com/mcp/setup/chatgpt) | | Claude Code | Run `claude mcp add --transport http --scope user 3daistudio https://mcp.3daistudio.com/mcp`. [Full guide](https://docs.3daistudio.com/mcp/setup/claude-code) | | Cursor | Add the server to your `mcp.json`. [Full guide](https://docs.3daistudio.com/mcp/setup/cursor) | | Codex | Run `codex mcp add 3daistudio --url https://mcp.3daistudio.com/mcp`. [Full guide](https://docs.3daistudio.com/mcp/setup/codex) | | VS Code (Copilot) | Add the server to `.vscode/mcp.json`. [Full guide](https://docs.3daistudio.com/mcp/setup/vs-code) | | OpenCode | Add the server to `opencode.json`. [Full guide](https://docs.3daistudio.com/mcp/setup/opencode) | | Anything else | Add a remote (Streamable HTTP) MCP server with the URL. [Full guide](https://docs.3daistudio.com/mcp/setup/other-clients) | **Menus or commands look different?** AI apps change their MCP settings often, so these steps can go out of date within days. Your assistant usually knows its current setup, so just ask it. Coding agents can often set it up for you: ```text How do I add a remote MCP server in this app right now? I want to connect 3D AI Studio: https://mcp.3daistudio.com/mcp (OAuth sign in, no API key). Use https://docs.3daistudio.com/mcp-setup.md for the details, and set it up for me if you can. ``` ### Step 3: Sign in and approve Your assistant opens a 3D AI Studio page in your browser. 1. **Sign in** with your 3D AI Studio account (skip this if you are already signed in). 2. **Review the permissions** under "This app will be able to". The page lists exactly what the assistant can do. See [Security & Privacy](https://docs.3daistudio.com/mcp/security-privacy) for what each one means. 3. **Choose which credits to use** in the "Use credits from" menu. **Personal** is selected by default. If you belong to a team, you can pick the team's credits instead. 4. Click **Authorize**. The page tells you which app is asking to connect. If you did not start a connection yourself, click **Cancel**. ### Step 4: Generate your first model Go back to your assistant and ask for something. A few ideas: - "Make a low poly fox for my game." - "Create a three headed wolf statue I can 3D print, and export it as an STL." - "Generate an image of a cozy wizard hut, then turn it into a 3D model." - "What 3D models can you make with 3D AI Studio, and what do they cost?" The assistant starts the generation and checks on it until it is done. 3D models usually take **1 to 5 minutes**; images take seconds. The result appears in the chat (or as a link, depending on your assistant) and in your [dashboard](https://www.3daistudio.com/Dashboard). ## Check that it worked Ask your assistant: **"What's my 3D AI Studio credit balance?"** If it answers with a number, you are connected. If it says it has no such tool, or asks you to sign in again, see [Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting). You can also open [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp). Your assistant should be listed under **Connected Apps**. ## Step by step guides for each assistant - [Claude](https://docs.3daistudio.com/mcp/setup/claude) - web and desktop, with the interactive 3D viewer in the chat - [ChatGPT](https://docs.3daistudio.com/mcp/setup/chatgpt) - web and desktop - [Claude Code](https://docs.3daistudio.com/mcp/setup/claude-code) - terminal - [Cursor](https://docs.3daistudio.com/mcp/setup/cursor) - editor - [Codex](https://docs.3daistudio.com/mcp/setup/codex) - CLI, IDE extension and app - [VS Code (Copilot)](https://docs.3daistudio.com/mcp/setup/vs-code) - editor - [OpenCode](https://docs.3daistudio.com/mcp/setup/opencode) - terminal - [Other MCP clients](https://docs.3daistudio.com/mcp/setup/other-clients) - anything that supports remote MCP servers > [!NOTE] > **Using a coding agent?** Claude Code, Codex, Cursor and similar tools can set up the connector themselves. Send them the [agent setup page](https://docs.3daistudio.com/mcp/agent-setup). ## Frequently Asked Questions ### How long does setup take? About two minutes: paste one URL, sign in, click Authorize. ### Do I have to set it up again on every device? For chat apps like Claude and ChatGPT, the connector is tied to your assistant account, so it is usually available wherever you use that account. Developer tools like Claude Code, Codex and Cursor store their configuration per machine (or per project), so you add it once on each computer. ### Can I connect more than one assistant? Yes. Each assistant is its own connection, and you can see and disconnect each one separately in [Managing Connections](https://docs.3daistudio.com/mcp/managing-connections). ### What if I am on the free plan? You can add the connector, but the sign in page will ask you to upgrade before access is granted. After upgrading, simply connect again. --- # Set Up With Your Agent Source: https://docs.3daistudio.com/mcp/agent-setup Coding agents like **Claude Code, Codex, Cursor, VS Code Copilot and OpenCode** can add the 3D AI Studio connector to their own configuration. You only have to do one thing yourself: sign in and click **Authorize** when the browser opens. This is often the most reliable way to connect. AI apps change their MCP commands and settings frequently, sometimes within days, while the agent itself knows how its current version works. The instructions below give it everything else it needs. There are two ways to hand the job to your agent. Use whichever is easier. ## Option 1: Point your agent at the setup file Paste this into your agent: ```text Set up the 3D AI Studio MCP server by following https://docs.3daistudio.com/mcp-setup.md ``` The file at [docs.3daistudio.com/mcp-setup.md](https://docs.3daistudio.com/mcp-setup.md) is plain Markdown written for AI agents. It contains the exact configuration for each client, how to start the sign in, how to verify the connection, and what to do if something goes wrong. Your agent needs web access to read it. ## Option 2: Paste the brief If your agent cannot open links, copy this whole block into the chat instead: ```text Please connect the 3D AI Studio MCP server for me. Server name: 3daistudio URL: https://mcp.3daistudio.com/mcp Transport: remote MCP server over Streamable HTTP Authentication: OAuth sign in in my browser. No API key, no headers, no env vars. 1. Add the server to the client you are running in: - Claude Code: claude mcp add --transport http --scope user 3daistudio https://mcp.3daistudio.com/mcp - Codex: codex mcp add 3daistudio --url https://mcp.3daistudio.com/mcp - Cursor: add "3daistudio": { "url": "https://mcp.3daistudio.com/mcp" } under "mcpServers" in ~/.cursor/mcp.json - VS Code Copilot: add "3daistudio": { "type": "http", "url": "https://mcp.3daistudio.com/mcp" } under "servers" in .vscode/mcp.json - OpenCode: add "3daistudio": { "type": "remote", "url": "https://mcp.3daistudio.com/mcp", "enabled": true } under "mcp" in opencode.json Merge into existing config files. Do not remove other servers. 2. Start the sign in (Claude Code: I run /mcp and choose Authenticate; Codex: codex mcp login 3daistudio; OpenCode: opencode mcp auth 3daistudio). I will sign in and click Authorize in the browser myself. 3. Never write tokens or secrets into config files. 4. When I confirm I am signed in, call get_credit_balance to verify the connection and tell me the result. Full instructions: https://docs.3daistudio.com/mcp-setup.md ``` ## What your agent will do 1. **Add the server** to its configuration, using the server name `3daistudio` and the URL `https://mcp.3daistudio.com/mcp`. 2. **Start the sign in.** A 3D AI Studio page opens in your browser. 3. **Wait for you.** You sign in, review the permissions, choose Personal or team credits, and click **Authorize**. 4. **Verify** by asking 3D AI Studio for your credit balance. After that, you can ask for 3D models and images directly in the same agent session (some clients need a restart or a reload of their MCP servers first). > [!WARNING] > **Only you can authorize the connection.** No agent can complete the sign in for you, and none should ever ask for your 3D AI Studio password or put tokens into a config file. If an agent asks for either, stop and check what it is doing. ## Chat assistants (Claude, ChatGPT) Chat assistants cannot change their own settings, so you add the connector yourself. It takes about a minute: - **Claude:** Settings → Connectors → Add custom connector → paste `https://mcp.3daistudio.com/mcp`. [Full guide](https://docs.3daistudio.com/mcp/setup/claude) - **ChatGPT:** Settings → Connectors → Create → paste `https://mcp.3daistudio.com/mcp`. [Full guide](https://docs.3daistudio.com/mcp/setup/chatgpt) ## For AI tools and crawlers - [`/mcp-setup.md`](https://docs.3daistudio.com/mcp-setup.md) - agent setup instructions (this page in plain Markdown, with full client configs) - [`/llms.txt`](https://docs.3daistudio.com/llms.txt) - index of the whole documentation for language models - [`/llms-full.txt`](https://docs.3daistudio.com/llms-full.txt) - the documentation in a single Markdown file - Any page as Markdown: add `.md` to its address, for example [`/mcp/quick-start.md`](https://docs.3daistudio.com/mcp/quick-start.md). Every page also has an **Ask AI** menu (top right) with Copy Markdown and View as Markdown ## Frequently Asked Questions ### Is it safe to let an agent set this up? Yes. The agent only adds a server address to its own configuration. Access to your account is granted by you, in your browser, on the 3D AI Studio sign in page. The agent never sees your password, and you can disconnect it at any time in [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp). ### My agent says it cannot find the tools after setup. What now? Most clients load MCP servers at startup. Restart the agent (or reload its MCP servers), then ask it to call `get_credit_balance`. If it still fails, see [Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting). ### Can the agent set it up on a remote machine or in a container? The sign in needs a browser. If your agent runs somewhere without one, set up the connector on a machine where you can open the sign in page, or use your client's own guidance for authenticating remote servers on headless machines. --- # Add 3D AI Studio to Claude Connectors Source: https://docs.3daistudio.com/mcp/setup/claude Claude gives you the best experience with 3D AI Studio: finished models open in an **interactive 3D viewer right in the conversation**, and you can drop images into an inline upload card. This guide covers Claude on the web (claude.ai) and the Claude desktop app. New to 3D AI Studio? See how to [generate 3D models in Claude](https://www.3daistudio.com/MCP) with the 3D model generation MCP. Using Claude in the terminal? See [Claude Code](https://docs.3daistudio.com/mcp/setup/claude-code) instead. ## What you need - A **3D AI Studio account on a paid plan**. - A **Claude account that can add custom connectors.** Custom connectors depend on your Claude plan. In Claude Team and Enterprise workspaces, an owner may need to add the connector for the organization first (see [below](https://docs.3daistudio.com/mcp/setup/claude#claude-team-and-enterprise)). - The connector URL: `https://mcp.3daistudio.com/mcp` > [!NOTE] > **Menus look different?** Claude updates its settings often, so the exact menu names below may already have changed. Claude itself usually knows the current steps. Ask it in a chat (turn on web search if you have it): ```text How do I add a custom connector (remote MCP server) in Claude right now? I want to connect 3D AI Studio: https://mcp.3daistudio.com/mcp (OAuth sign in, no API key). Use https://docs.3daistudio.com/mcp-setup.md for the details and walk me through the exact steps for my plan. ``` ## Setup ### Open the connector settings In Claude on the web or in the desktop app, click your profile and open **Settings → Connectors**. ### Add a custom connector Click **Add custom connector** and fill in: | Field | Value | | --- | --- | | Name | `3D AI Studio` | | Remote MCP server URL | `https://mcp.3daistudio.com/mcp` | Leave any advanced or OAuth fields empty. Click **Add**. ### Connect and authorize Click **Connect** next to 3D AI Studio. A 3D AI Studio page opens: 1. Sign in with your 3D AI Studio account. 2. Review what Claude will be able to do. 3. Under **Use credits from**, keep **Personal** or choose a team. 4. Click **Authorize**. You are sent back to Claude, and the connector shows as connected. ### Use it in a chat Start a new chat. Open the tools menu in the message box (the "Search and tools" or connectors menu) and make sure **3D AI Studio** is turned on. Then ask: > "Make a low poly fox for my game." ## Check that it worked Ask Claude: **"What's my 3D AI Studio credit balance?"** Claude should call the connector and answer with a number. The connection also appears under **Connected Apps** in [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp) on 3D AI Studio. ## What it looks like in Claude ![A finished 3D model in the 3D AI Studio viewer inside a Claude conversation, with Download and Open in 3D AI Studio buttons](https://docs.3daistudio.com/assets/images/docs/mcp/mcp-chat-viewer.webp) - **3D viewer in the chat.** When a model is finished, Claude shows it in an interactive viewer. You can spin it, switch between the textured and clay (untextured) view, and open it in your dashboard. - **Image cards.** Generated and edited images appear as cards in the conversation. - **Upload card.** When you want to use your own photo or sketch, Claude opens an upload card you can drop the file into. See [Images & Uploads](https://docs.3daistudio.com/mcp/images-and-uploads). - **Progress.** 3D models take 1 to 5 minutes. Claude checks on the generation by itself until it is done, so you can simply wait. ## Claude Team and Enterprise In organization workspaces, custom connectors are usually managed by an owner: 1. An owner adds the connector once in the organization's admin connector settings, using the URL `https://mcp.3daistudio.com/mcp`. 2. Each member then goes to **Settings → Connectors**, clicks **Connect** next to 3D AI Studio, and signs in with **their own** 3D AI Studio account. Every member authorizes separately, so each person's generations use the credits they choose on the sign in page (personal or a team they belong to). ## Claude specific fixes | Problem | Fix | | --- | --- | | No **Add custom connector** button | Your Claude plan or workspace does not allow custom connectors, or an owner has to add it for your organization | | Claude answers without using 3D AI Studio | Turn on 3D AI Studio in the chat's tools menu, or ask explicitly: "use 3D AI Studio to make..." | | "3D AI Studio is signed out" or a reconnect prompt | Open **Settings → Connectors**, click 3D AI Studio and connect again | | The connector fails right after you paste the URL | Check it is exactly `https://mcp.3daistudio.com/mcp` (not `www.`, not ending in `/sse`) | More fixes are on the [Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting) page. ## Frequently Asked Questions ### Does it work in the Claude desktop app? Yes. Connectors are tied to your Claude account, so once you add 3D AI Studio on claude.ai it is also available in the desktop app, and the 3D viewer works there too. ### Can I see the 3D model without leaving Claude? Yes. Claude supports interactive cards, so finished models open in a 3D viewer inside the conversation. You can still open every model in your 3D AI Studio dashboard to download it in other formats or edit it. ### Does Claude spend my credits without asking? Claude only generates when you ask it to, and every generation costs the same credits as on the website. Very large requests ask you to confirm first. You can always ask "how much will this cost?" before generating. See [Credits & Limits](https://docs.3daistudio.com/mcp/credits-and-limits). --- # Connect ChatGPT (Connectors & MCP) Source: https://docs.3daistudio.com/mcp/setup/chatgpt ChatGPT cannot create 3D models on its own. With the 3D AI Studio connector added, it can: you ask for a model in the chat, the generation runs on 3D AI Studio, and a real textured mesh with downloadable files comes back. New to 3D AI Studio? See how to [generate 3D models in ChatGPT](https://www.3daistudio.com/MCP) with the 3D model generation MCP. ## What you need - A **3D AI Studio account on a paid plan**. - A **ChatGPT plan that supports custom connectors.** Availability depends on your ChatGPT plan and, for Business and Enterprise workspaces, on what your admin allows. - The connector URL: `https://mcp.3daistudio.com/mcp` > [!NOTE] > **Menus look different?** ChatGPT renames and moves its connector settings often, so the steps below can go out of date quickly. Look for **Connectors**, **Apps**, or **Developer mode** in Settings, or simply ask ChatGPT itself. It usually knows the current way (turn on web search): ```text How do I add a custom MCP connector in ChatGPT right now? I want to connect 3D AI Studio: https://mcp.3daistudio.com/mcp (OAuth sign in, no API key). Use https://docs.3daistudio.com/mcp-setup.md for the details and walk me through the exact steps for my plan. ``` ## Setup ### Open the connector settings In ChatGPT on the web or desktop, open **Settings → Connectors** (in some versions: **Apps & Connectors**). If you do not see a **Create** button, open **Advanced settings** in the same section and turn on **Developer mode**. ChatGPT requires it for custom connectors on some plans. ### Create the connector Click **Create** and fill in: | Field | Value | | --- | --- | | Name | `3D AI Studio` | | Description | `Generate 3D models and images` (optional) | | MCP server URL | `https://mcp.3daistudio.com/mcp` | | Authentication | `OAuth` | Confirm that you trust the application if ChatGPT asks, then click **Create**. ### Sign in and authorize A 3D AI Studio page opens: 1. Sign in with your 3D AI Studio account. 2. Review what ChatGPT will be able to do. 3. Under **Use credits from**, keep **Personal** or choose a team. 4. Click **Authorize**. ### Use it in a chat Start a new chat, open the **+** or tools menu in the message box, and select **3D AI Studio** (with Developer mode, it is listed under the developer tools). Then ask: > "Make a three headed wolf statue I can 3D print." ## Check that it worked Ask ChatGPT: **"Use 3D AI Studio to check my credit balance."** It should call the connector and answer with a number. The connection also appears under **Connected Apps** in [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp) on 3D AI Studio. ## What it looks like in ChatGPT > [!NOTE] > **Name the tool in your request.** Say "**use 3D AI Studio** to make this 3D". Otherwise ChatGPT tends to reach for its own built in image generator first. - ChatGPT starts the generation and checks on it until it is finished. 3D models take 1 to 5 minutes; images take seconds. - Depending on your ChatGPT version, the finished result shows as a card or as a link. Either way, the model is saved to your [dashboard](https://www.3daistudio.com/Dashboard), where you can view it in 3D and download it in any format. - To use your own image, ChatGPT gives you an upload card or a link to an upload page. See [Images & Uploads](https://docs.3daistudio.com/mcp/images-and-uploads). ## ChatGPT specific fixes | Problem | Fix | | --- | --- | | No **Create** button under Connectors | Turn on **Developer mode** under Advanced settings, or check that your ChatGPT plan or workspace allows custom connectors | | ChatGPT answers without using 3D AI Studio | Select 3D AI Studio from the tools menu in the chat, or say "use 3D AI Studio to..." | | ChatGPT asks you to connect again | Open **Settings → Connectors**, select 3D AI Studio and connect again | | Creating the connector fails | Check the URL is exactly `https://mcp.3daistudio.com/mcp` and that authentication is set to OAuth | More fixes are on the [Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting) page. ## Frequently Asked Questions ### Can ChatGPT generate 3D models? Not by itself. ChatGPT has no 3D generation engine, so on its own it can at best write code for simple shapes. Connected to 3D AI Studio, it can generate real, textured 3D models with downloadable GLB, FBX, OBJ, STL, USDZ and 3MF files. For the full comparison of what works, see [Can ChatGPT create 3D models?](https://www.3daistudio.com/blog/chatgpt-3d-models-ai-powered-alternatives) ### Do I need an API key for ChatGPT? No. ChatGPT signs in to 3D AI Studio through a normal login page (OAuth). No key is created or stored. ### Is it the same as using the website? It uses the same AI models, the same credits and the same dashboard as the website. The difference is that you ask for things in your ChatGPT conversation instead of using the website's tools. --- # Claude Code MCP Setup Source: https://docs.3daistudio.com/mcp/setup/claude-code Claude Code can generate 3D models, images and materials for your project straight from the terminal. Ask for a game asset, a texture set or a 3D printable part, and Claude Code creates it on 3D AI Studio and gives you the result. New to 3D AI Studio? See the [3D model generation MCP](https://www.3daistudio.com/MCP) for what it can do and example results. > [!NOTE] > Want Claude Code to do the setup itself? Paste the brief from [Set Up With Your Agent](https://docs.3daistudio.com/mcp/agent-setup). ## What you need - A **3D AI Studio account on a paid plan**. - **Claude Code**, updated to a recent version. - A browser on the same machine for the one time sign in. > [!NOTE] > **Commands not working?** Claude Code changes how it handles MCP servers from time to time, so the commands below can go out of date. The quickest fix is to let Claude Code do it. Paste this into a session: ```text Add the 3D AI Studio MCP server for me: https://mcp.3daistudio.com/mcp (remote server, Streamable HTTP, OAuth sign in, no API key). Follow https://docs.3daistudio.com/mcp-setup.md, and if your version works differently, use the current way to add a remote MCP server. ``` ## Setup ### Add the server Run this in your terminal: ```bash claude mcp add --transport http --scope user 3daistudio https://mcp.3daistudio.com/mcp ``` - `--scope user` makes 3D AI Studio available in all your projects. Leave it out to add it to the current project only. - `3daistudio` is the server name. You can pick another name, but the commands on this page assume `3daistudio`. ### Authenticate Start Claude Code and run: ```text /mcp ``` Select **3daistudio** and choose **Authenticate**. A 3D AI Studio page opens in your browser: 1. Sign in with your 3D AI Studio account. 2. Under **Use credits from**, keep **Personal** or choose a team. 3. Click **Authorize**. Return to Claude Code. The server now shows as connected in `/mcp`. ### Try it Ask Claude Code: > "Use 3D AI Studio to make a low poly treasure chest for my game and give me the GLB." ## Check that it worked - Run `claude mcp list` in your terminal. `3daistudio` should be listed. - Inside Claude Code, ask: **"What's my 3D AI Studio credit balance?"** ## Working with results in the terminal - Claude Code shows finished models and images as **links** (there is no inline 3D viewer in the terminal). Open them in your browser or in your [dashboard](https://www.3daistudio.com/Dashboard). - To use a local image (a sketch, a photo, a concept), Claude Code gives you a **link to an upload page**. Open it, drop the file in, then tell Claude Code you are done. See [Images & Uploads](https://docs.3daistudio.com/mcp/images-and-uploads). - Ask Claude Code to **convert** a model to the format your project needs (for example "convert it to FBX" or "export an STL scaled to 10 cm"). ## Claude Code specific fixes | Problem | Fix | | --- | --- | | `3daistudio` shows as needing authentication | Run `/mcp`, select `3daistudio` and choose **Authenticate** | | The tools do not appear | Restart Claude Code, then check `claude mcp list`. Make sure the URL is exactly `https://mcp.3daistudio.com/mcp` | | A session says it has an outdated sign in | Close all running Claude Code sessions and start a new one. Your connection itself is fine | | You want to remove the server | Run `claude mcp remove 3daistudio`, and disconnect it in [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp) | More fixes are on the [Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting) page. ## Frequently Asked Questions ### Does Claude Code need an API key for 3D AI Studio? No. Claude Code signs in through your browser (OAuth). Do not add any API key or token to the configuration. ### Can I share the setup with my team through the repository? You can add the server at project scope so the configuration is shared, but each teammate still signs in with their own 3D AI Studio account. Sign ins are never shared through the repository. ### Where are the generated files? In your 3D AI Studio dashboard, and Claude Code gives you the download links in the conversation. Ask it to save a file into your project if you want it on disk. --- # Cursor MCP Server Setup Source: https://docs.3daistudio.com/mcp/setup/cursor With 3D AI Studio connected, Cursor's agent can create 3D models, images and PBR materials for the project you are working on: placeholder art for a game prototype, a product model for a web viewer, or textures for a scene. New to 3D AI Studio? See the [3D model generation MCP](https://www.3daistudio.com/MCP) for what it can do and example results. > [!NOTE] > Want Cursor to do the setup itself? Paste the brief from [Set Up With Your Agent](https://docs.3daistudio.com/mcp/agent-setup) into Cursor's agent. ## What you need - A **3D AI Studio account on a paid plan**. - **Cursor**, updated to a recent version. > [!NOTE] > **Settings look different?** Cursor moves its MCP settings around between versions, so the steps below can go out of date. Cursor's agent can usually set it up for you. Paste this into the agent chat: ```text Add the 3D AI Studio MCP server to my Cursor setup: https://mcp.3daistudio.com/mcp (remote server, OAuth sign in, no API key). Follow https://docs.3daistudio.com/mcp-setup.md, and if this Cursor version works differently, use the current way to add a remote MCP server. ``` ## Setup ### Add the server Open (or create) one of these files: - `~/.cursor/mcp.json` to make 3D AI Studio available in **all** projects, or - `.cursor/mcp.json` in a project folder to use it in **that project only**. Add the `3daistudio` entry. If the file already has other servers, add it next to them inside `mcpServers`: ```json { "mcpServers": { "3daistudio": { "url": "https://mcp.3daistudio.com/mcp" } } } ``` Do not add any API key, header or token. Sign in happens in the browser. ### Sign in Open **Cursor Settings** and go to the **MCP** section (called **Tools & MCP** in recent versions). Find **3daistudio** in the list. If it shows a **login**, **connect** or **needs authentication** action, click it. A 3D AI Studio page opens: 1. Sign in with your 3D AI Studio account. 2. Under **Use credits from**, keep **Personal** or choose a team. 3. Click **Authorize**. Back in Cursor, the server turns active and lists its tools. ### Try it Open the agent chat and ask: > "Use 3D AI Studio to generate a stylized wooden crate as a GLB for my game." ## Check that it worked In **Cursor Settings → MCP**, `3daistudio` should be enabled with its tools listed. Then ask the agent: **"What's my 3D AI Studio credit balance?"** ## Tips for using it in Cursor - Cursor shows results as **links**. Open them in the browser or in your [dashboard](https://www.3daistudio.com/Dashboard). Ask the agent to download a file into your project if you want it in the repository. - To use a local image, the agent gives you a **link to an upload page**. See [Images & Uploads](https://docs.3daistudio.com/mcp/images-and-uploads). - You can turn individual tools on or off in Cursor's MCP settings, for example to allow only image tools in a project. ## Cursor specific fixes | Problem | Fix | | --- | --- | | `3daistudio` shows an error or stays red | Check the JSON is valid and the URL is exactly `https://mcp.3daistudio.com/mcp`. Then toggle the server off and on | | The server asks you to log in again | Click its login or connect action in Cursor Settings → MCP and authorize again | | The agent does not use the tools | Make sure you are using the agent (not a plain chat mode) and that `3daistudio` is enabled | More fixes are on the [Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting) page. ## Frequently Asked Questions ### Should I commit `.cursor/mcp.json` to my repository? You can. It only contains the server address, no credentials. Every teammate signs in with their own 3D AI Studio account. ### Do I need an API key? No. Cursor signs in through your browser. Never put tokens into `mcp.json`. --- # Codex MCP Setup Source: https://docs.3daistudio.com/mcp/setup/codex Codex can generate 3D models, images and materials while it works on your code. The **Codex CLI, the IDE extension and the Codex app share one configuration**, so you set up 3D AI Studio once and it works in all three. New to 3D AI Studio? See the [3D model generation MCP](https://www.3daistudio.com/MCP) for what it can do and example results. > [!NOTE] > Want Codex to do the setup itself? Paste the brief from [Set Up With Your Agent](https://docs.3daistudio.com/mcp/agent-setup). ## What you need - A **3D AI Studio account on a paid plan**. - **Codex**, updated to a recent version (remote MCP servers with sign in need a current release). - A browser on the same machine for the sign in. > [!NOTE] > **Commands not working?** Codex changes its MCP commands from time to time, so the steps below can go out of date. The quickest fix is to let Codex do it. Paste this into a session: ```text Add the 3D AI Studio MCP server for me: https://mcp.3daistudio.com/mcp (remote server, Streamable HTTP, OAuth sign in, no API key). Follow https://docs.3daistudio.com/mcp-setup.md, and if your version works differently, use the current way to add a remote MCP server. ``` ## Setup ### Add the server Run: ```bash codex mcp add 3daistudio --url https://mcp.3daistudio.com/mcp ``` If you prefer editing the configuration file, add this to `~/.codex/config.toml` instead: ```toml [mcp_servers.3daistudio] url = "https://mcp.3daistudio.com/mcp" ``` ### Sign in Codex opens a 3D AI Studio sign in page in your browser. If it does not, run: ```bash codex mcp login 3daistudio ``` On the 3D AI Studio page: 1. Sign in with your 3D AI Studio account. 2. Under **Use credits from**, keep **Personal** or choose a team. 3. Click **Authorize**. ### Try it Start a new Codex session and ask: > "Use 3D AI Studio to make a low poly fox and give me the GLB link." ## Check that it worked - Run `codex mcp list`. `3daistudio` should be listed. - In a new session, ask: **"What's my 3D AI Studio credit balance?"** ## Keep the sign in healthy Codex can run several sessions (terminals, editor windows, the app) on the same sign in. A few habits avoid sign in trouble: - **If you are ever signed out**, run `codex mcp login 3daistudio` and authorize again. - **If 3D AI Studio shows "Restart needed"** in [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp), one of your Codex sessions is still using an old sign in. Close all running Codex sessions and start a new one. Your connection itself is fine. - **Update Codex** regularly. Newer versions handle sign ins across several sessions more reliably. ## Working with results - Codex shows finished models and images as **links**. Open them in the browser or in your [dashboard](https://www.3daistudio.com/Dashboard), or ask Codex to download them into your project. - To use a local image, Codex gives you a **link to an upload page**. See [Images & Uploads](https://docs.3daistudio.com/mcp/images-and-uploads). ## Codex specific fixes | Problem | Fix | | --- | --- | | Codex keeps saying the server needs authentication | Run `codex mcp login 3daistudio`. If it keeps happening, close all Codex sessions, update Codex, and log in again | | `codex mcp add` does not accept `--url` | Your Codex version is too old for remote servers. Update Codex | | The tools do not appear in a session | Start a new session after adding the server. Check `codex mcp list` | | You used another server name | Replace `3daistudio` in the commands with your name, for example `codex mcp login my-3d-server` | More fixes are on the [Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting) page. ## Frequently Asked Questions ### Do the Codex CLI, IDE extension and app need separate setups? No. They share the same Codex configuration, so adding 3D AI Studio once makes it available in all of them. ### Do I need an API key? No. Codex signs in through your browser. Do not put tokens into `config.toml`. --- # VS Code MCP Setup (GitHub Copilot) Source: https://docs.3daistudio.com/mcp/setup/vs-code GitHub Copilot's agent mode in VS Code can use the 3D AI Studio tools to create 3D models, images and materials for your project. New to 3D AI Studio? See the [3D model generation MCP](https://www.3daistudio.com/MCP) for what it can do and example results. ## What you need - A **3D AI Studio account on a paid plan**. - **VS Code** with **GitHub Copilot**, both up to date, and MCP support enabled (in some organizations an admin has to allow MCP servers). > [!NOTE] > **Steps look different?** VS Code and Copilot change their MCP support frequently, so the steps below can go out of date. Copilot's agent mode can usually set it up for you. Paste this into Copilot Chat in Agent mode: ```text Add the 3D AI Studio MCP server to VS Code for me: https://mcp.3daistudio.com/mcp (remote server, type http, OAuth sign in, no API key). Follow https://docs.3daistudio.com/mcp-setup.md, and if this version works differently, use the current way to add a remote MCP server. ``` ## Setup ### Add the server **Option A: use the command.** Open the Command Palette and run **MCP: Add Server**. Choose **HTTP**, paste `https://mcp.3daistudio.com/mcp`, name it `3daistudio`, and choose whether to save it for this workspace or globally. **Option B: edit the file.** Create or open `.vscode/mcp.json` in your workspace and add: ```json { "servers": { "3daistudio": { "type": "http", "url": "https://mcp.3daistudio.com/mcp" } } } ``` If the file already lists other servers, add `3daistudio` next to them inside `servers`. ### Start the server and sign in VS Code shows a **Start** action above the server entry in `mcp.json` (or run **MCP: List Servers**, select `3daistudio` and start it). When VS Code asks to authenticate, allow it. A 3D AI Studio page opens: 1. Sign in with your 3D AI Studio account. 2. Under **Use credits from**, keep **Personal** or choose a team. 3. Click **Authorize**. ### Use it in agent mode Open **Copilot Chat**, switch to **Agent** mode, and check in the tools picker that the 3daistudio tools are enabled. Then ask: > "Use 3D AI Studio to generate a sci-fi crate as a GLB for my scene." ## Check that it worked Run **MCP: List Servers**. `3daistudio` should be running. Then ask Copilot: **"What's my 3D AI Studio credit balance?"** ## VS Code specific fixes | Problem | Fix | | --- | --- | | The server does not start | Check `mcp.json` is valid JSON, `type` is `http` and the URL is exactly `https://mcp.3daistudio.com/mcp` | | Copilot does not use the tools | Use **Agent** mode and enable the 3daistudio tools in the tools picker | | You are asked to sign in again | Restart the server from **MCP: List Servers** and accept the authentication prompt | | MCP options are missing | Update VS Code and Copilot. In managed setups, ask your admin whether MCP servers are allowed | More fixes are on the [Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting) page. ## Frequently Asked Questions ### Can I share `.vscode/mcp.json` with my team? Yes. It contains only the server address. Each teammate signs in with their own 3D AI Studio account. ### Do I need an API key or input variable? No. Do not add headers, tokens or `inputs` for 3D AI Studio. Sign in happens in the browser. --- # OpenCode MCP Server Setup Source: https://docs.3daistudio.com/mcp/setup/opencode OpenCode can use 3D AI Studio as a remote MCP server to create 3D models, images and materials from the terminal. New to 3D AI Studio? See the [3D model generation MCP](https://www.3daistudio.com/MCP) for what it can do and example results. ## What you need - A **3D AI Studio account on a paid plan**. - **OpenCode**, updated to a recent version. - A browser on the same machine for the sign in. > [!NOTE] > **Config not working?** OpenCode updates its MCP configuration from time to time, so the steps below can go out of date. OpenCode can usually set it up for you. Paste this into a session: ```text Add the 3D AI Studio MCP server for me: https://mcp.3daistudio.com/mcp (remote server, OAuth sign in, no API key). Follow https://docs.3daistudio.com/mcp-setup.md, and if your version works differently, use the current way to add a remote MCP server. ``` ## Setup ### Add the server Add a remote MCP server named `3daistudio` to your `opencode.json` (in the project, or in your global OpenCode config). If the file already has an `mcp` section, add the entry inside it: ```json { "$schema": "https://opencode.ai/config.json", "mcp": { "3daistudio": { "type": "remote", "url": "https://mcp.3daistudio.com/mcp", "enabled": true } } } ``` > [!WARNING] > Use the URL exactly as shown. `https://www.3daistudio.com` does **not** work as an MCP server address. ### Sign in Run: ```bash opencode mcp auth 3daistudio ``` A 3D AI Studio page opens: 1. Sign in with your 3D AI Studio account. 2. Under **Use credits from**, keep **Personal** or choose a team. 3. Click **Authorize**. ### Try it Start OpenCode and ask: > "Use 3D AI Studio to make a low poly tree as a GLB." ## Check that it worked Ask OpenCode: **"What's my 3D AI Studio credit balance?"** It should call the connector and answer with a number. ## OpenCode specific fixes | Problem | Fix | | --- | --- | | Signed out, or the server needs authentication | Run `opencode mcp auth 3daistudio` again | | "Restart needed" in 3D AI Studio settings | Close all running OpenCode sessions and start a new one. Your connection itself is fine | | The server fails to connect | Check the URL is exactly `https://mcp.3daistudio.com/mcp` and `type` is `remote` | | You used another server name | Replace `3daistudio` in the commands with your name | Results show as **links** in the terminal, and image uploads use a browser upload page. See [Images & Uploads](https://docs.3daistudio.com/mcp/images-and-uploads). More fixes are on the [Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting) page. --- # Add 3D AI Studio to Any MCP Client Source: https://docs.3daistudio.com/mcp/setup/other-clients 3D AI Studio works with any MCP client that supports **remote MCP servers with sign in**. If your client is not covered by a dedicated guide, use the details on this page. New to 3D AI Studio? See the [3D model generation MCP](https://www.3daistudio.com/MCP) for what it can do and example results. ## Connection details | Setting | Value | | --- | --- | | Server URL | `https://mcp.3daistudio.com/mcp` | | Suggested server name | `3daistudio` | | Transport | Remote MCP server over **Streamable HTTP** (often labeled "http", "remote" or "streamable-http") | | Authentication | **OAuth**, through the standard MCP authorization flow. Your client registers itself automatically and opens a browser sign in | | API key, headers, environment variables | **None.** Do not add any | The sign in shows a 3D AI Studio page where you sign in, choose which credits to use (Personal or a team) and click **Authorize**. > [!WARNING] > Use the URL exactly as shown. These do **not** work: `https://www.3daistudio.com`, `https://mcp.3daistudio.com` without `/mcp`, and older guides that end in `/sse`. > [!NOTE] > **Not sure where your client puts MCP servers?** Ask your assistant directly. Most assistants know how their own app adds remote MCP servers, and coding agents can often do the setup for you: ```text How do I add a remote MCP server in this app? I want to connect 3D AI Studio: https://mcp.3daistudio.com/mcp (Streamable HTTP, OAuth sign in, no API key). Use https://docs.3daistudio.com/mcp-setup.md for the details, and set it up for me if you can. ``` ## Common configuration shapes Every client names its settings a little differently. These are the most common shapes. Check your client's documentation for the exact field names. **`url` style** (Cursor and many others): ```json { "mcpServers": { "3daistudio": { "url": "https://mcp.3daistudio.com/mcp" } } } ``` **`type` + `url` style** (VS Code and others): ```json { "servers": { "3daistudio": { "type": "http", "url": "https://mcp.3daistudio.com/mcp" } } } ``` Some clients call the field `serverUrl` or `httpUrl` instead of `url`. The value is always the same. ## Clients that only support local servers Some clients can only start local MCP servers (a command on your computer) and cannot connect to a remote URL. For those, you can use the community bridge [`mcp-remote`](https://www.npmjs.com/package/mcp-remote), which runs locally and forwards to the remote server, including the browser sign in: ```json { "mcpServers": { "3daistudio": { "command": "npx", "args": ["-y", "mcp-remote", "https://mcp.3daistudio.com/mcp"] } } } ``` This needs Node.js installed. `mcp-remote` is a third party tool and is not maintained by 3D AI Studio. Prefer your client's built in remote server support whenever it has one. ## What your client needs to support - **Streamable HTTP** remote servers. - **MCP authorization (OAuth)** with a browser sign in. - **Tools.** That is all you need to generate. - **Optional:** clients that support interactive app cards (like Claude) show finished models in a 3D viewer and offer an inline upload card. Other clients show links instead, which works just as well. ## Check that it worked Ask your client to call the `get_credit_balance` tool, or simply ask: **"What's my 3D AI Studio credit balance?"** If you get a number back, you are connected. If your client cannot connect, see [Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting). If you believe your client should work but does not, contact [support@3daistudio.com](mailto:support@3daistudio.com) and tell us the client name and version. --- # What You Can Do With the MCP Source: https://docs.3daistudio.com/mcp/capabilities This page lists everything your assistant can do once it is connected to 3D AI Studio, with the credit cost of each action and prompts you can copy. You never have to call tools by name. Just describe what you want, and the assistant picks the right tool. The tool names are listed so you can recognize them in your assistant's activity log, or ask for a specific one. > [!NOTE] > **Costs on this page are the standard prices at the time of writing.** Your assistant always has the current prices: ask "which 3D AI Studio models are available and what do they cost?" and it will check for you. Every action costs the same credits as on the website. ## At a glance | Category | What you can ask for | | --- | --- | | [3D generation](https://docs.3daistudio.com/mcp/capabilities#3d-generation) | 3D models from text, one image, or several views | | [Images](https://docs.3daistudio.com/mcp/capabilities#images) | Generate images from text, edit images with instructions | | [3D processing](https://docs.3daistudio.com/mcp/capabilities#3d-processing) | Low poly versions, UV unwrapping, format conversion and scaling | | [Materials](https://docs.3daistudio.com/mcp/capabilities#materials) | Tileable PBR materials from text or from a photo | | [Image utilities](https://docs.3daistudio.com/mcp/capabilities#image-utilities) | Background removal | | [Library and projects](https://docs.3daistudio.com/mcp/capabilities#library-and-projects) | Find past generations, show results, organize into projects | | [Uploads](https://docs.3daistudio.com/mcp/capabilities#uploads) | Hand your own photos and sketches to the assistant | | [Account](https://docs.3daistudio.com/mcp/capabilities#account) | Credit balance and the model catalog | ## 3D generation All 3D models support **text to 3D** (a description), **image to 3D** (one image), and **multi view to 3D** (several images of the same object from different sides). A 3D model usually takes **1 to 5 minutes**. Your assistant checks on it until it is done. | Model | Best for | Inputs | Typical cost | | --- | --- | --- | --- | | **Prism 3.1** (recommended) `generate_prism_3_1` | Best overall quality. The default choice for almost everything | Text, 1 image, or 2 to 4 views | 35 credits | | **P2** `generate_p2` | Low poly, game ready topology, with optional native **quad** mesh output | Text, 1 image, or 2 to 4 views | 70 credits with texture | | **P1** `generate_p1` | Previous generation low poly model, cheaper than P2 | Text, 1 image, or 2 to 4 views | 60 credits with texture | | **Hunyuan 3.1** `generate_hunyuan_3_1` | Clean, even meshes at a flat price. PBR included | Text, 1 image, or up to 8 views | 40 credits flat | | **Meshy 7.1** `generate_meshy_7` | Very high resolution 4K and 8K textures and ultra detailed geometry | 1 image, or 2 to 4 views (no text) | 40 textured, 35 mesh only | ### Options and surcharges You can ask for these in plain language ("with extreme texture detail", "as a quad mesh", "without texture"): | Option | Models | Extra cost | | --- | --- | --- | | Detailed texture / extreme texture | Prism 3.1 (+5 / +10), P2 and P1 (+10 / +15) | As listed | | Detailed geometry | Prism 3.1 | +10 | | Multi view input | Prism 3.1, P2, P1, Meshy 7.1, Hunyuan 3.1 | +5 | | No texture (geometry only) | P2 and P1 (saves the 20 credit texture charge), Meshy 7.1 (35 instead of 40) | Cheaper | | Quad mesh output | P2 | Free | | Newest texture model (experimental), with a fast texture option and lighting removal for photos | Prism 3.1 and P2 | Free | | 4K / 8K textures | Meshy 7.1 | +10 / +15 | | Ultra geometry | Meshy 7.1 | +10 (2K) / +15 (4K, single image) | | Low poly mode | Meshy 7.1 | 20 textured / 15 mesh only, plus texture surcharges | | Low poly, geometry only and sketch modes | Hunyuan 3.1 | Included in the flat price | > [!NOTE] > **For multi view input, order matters.** Give the views in the order front, left, back, right. One good image gives great results in most cases, so you only need multiple views when the back or sides of your object are very different from the front. **Example prompts** - "Make a low poly fox for my game." - "Turn this product photo into a 3D model with detailed textures." - "Create a three headed wolf statue I can 3D print." - "Generate a sci-fi crate with P2 as a quad mesh." - "Use Meshy with 8K textures on this image of a leather boot." ## Images Images take seconds and are a great first step: design the look as an image, then turn it into 3D. | Model | What it does | Cost per image | | --- | --- | --- | | **Nano Banana 2.1** (recommended for generating) `generate_image_nano_2_1` `edit_image_nano_2_1` | Fast, consistent images. Also edits with up to 14 input images | 6 (1K), 9 (2K), 13 (4K). Edits: +1 per started group of 3 input images. +3 with Google search grounding | | **GPT Image 2** (recommended for editing) `generate_image_gpt` `edit_image_gpt` | Very precise instruction following and text rendering. Edits up to 10 input images, optionally with a mask | Low quality 3 / 5 / 8, medium 6 / 14 / 25 (1K / 2K / 4K), high 30 | | **Nano Banana 2 Lite** `generate_image_nano_lite` | The cheapest image generation | 5 flat (+2 with Google search grounding) | | **Nano Banana 2** `edit_image_nano` | Previous generation editor | 7 (0.5K), 10 (1K), 14 (2K), 20 (4K). +2 with Google search grounding | You can generate 1 to 4 images per request. Generated images can be used directly as input for 3D generation. **Example prompts** - "Generate a cozy wizard hut on a white background, then turn it into a 3D model." - "Edit this photo: remove the person in the background." - "Make three variations of this character in different outfits." - "Generate a front view of a cartoon robot so I can turn it into 3D." ## 3D processing These work on a model you already have, either one you just generated or one from your library. Each produces a **new** model, so the original stays untouched. | Action | What it does | Cost | | --- | --- | --- | | **Low poly version** `low_poly_model` | Clean, lighter retopology for games, AR, VR and real time use. The default version (3.0) can unwrap UVs in the same step for free. Takes 1 to 3 minutes | 30 (35 as quad), 25 for version 2.0, 15 for version 1.0 (basic decimation) | | **UV unwrap** `unwrap_uvs` | Clean UV layout while keeping textures, rig and animation. For models up to 80,000 triangles (make a low poly version first for denser models). Also saves a free preview of the UV layout | 15 flat | | **Convert format** `convert_model` | GLB, OBJ, OBJ with textures (zip), FBX, FBX with textures (zip), STL, PLY, USDZ, 3MF. Can scale the model down to a maximum size in cm or inches, for 3D printing | 1 | **Example prompts** - "Make a low poly version of that model for my mobile game." - "Unwrap the UVs of my last model." - "Convert it to STL and scale it to 12 cm tall." - "Export the model as FBX with textures." ## Materials | Action | Output | Cost | | --- | --- | --- | | **Material from text** `generate_material` | A seamlessly tiling PBR material: base texture plus base color, normal, roughness, metalness and height maps | 10 for the whole set | | **Material from a photo** `material_from_image` | The same PBR maps extracted from a photo or texture image. Works best with flat, evenly lit close ups of a surface | 7 for the whole set | Materials are for **surfaces** (brick, rusted metal, moss, fabric) to use in a game engine or 3D app. They do not texture an existing 3D model. **Example prompts** - "Create a mossy cobblestone material for my level." - "Turn this photo of oak bark into a PBR material." ## Image utilities | Action | What it does | Cost | | --- | --- | --- | | **Remove background** `remove_background` | Keeps the subject on a transparent background. The cheapest option works for most images; the most expensive one is best for product photos | 1 to 3, depending on the quality option | You do **not** need to remove the background before turning an image into 3D. The 3D models handle backgrounds on their own. ## Library and projects | Action | Tool | Cost | | --- | --- | --- | | Browse and search your generations | `list_generations` | Free | | Check a generation's progress or result | `get_generation` | Free | | Show a finished result in the chat | `show_generation` | Free | | List your dashboard projects | `list_projects` | Free | | Create a project | `create_project` | Free | | Move a generation into a project | `assign_to_project` | Free | Everything your assistant creates lands in your [dashboard](https://www.3daistudio.com/Dashboard). You can also work with models you made on the website: "find my dragon model from last week and convert it to FBX". **Example prompts** - "Show me my last five 3D models." - "Create a project called Forest Level and put all models from today in it." - "Generate a barrel, a crate and a lantern, and put them in my Tavern project." ## Uploads | Action | Tool | Cost | | --- | --- | --- | | Hand an image from your device to 3D AI Studio | `upload_image` | Free | | List your recent uploads | `list_recent_uploads` | Free | Your assistant cannot pass images from the chat to 3D AI Studio directly. Instead, it opens an upload card in the chat (or gives you an upload link). See [Images & Uploads](https://docs.3daistudio.com/mcp/images-and-uploads) for how it works. ## Account | Action | Tool | Cost | | --- | --- | --- | | Check your credit balance | `get_credit_balance` | Free | | List models with their inputs and costs | `list_available_models` | Free | ## Example workflows **Game asset, from idea to engine** > "Generate a front view image of a stylized treasure chest, turn it into a 3D model with P2 as a quad mesh, then convert it to FBX with textures." **3D print** > "Turn this sketch into a 3D model, then convert it to STL scaled to 8 cm." **Product visualization** > "Make a 3D model of this sneaker photo with detailed textures and show it to me." **Level dressing** > "Create a project called Swamp, generate a mossy log, a lantern and a wooden sign, put them all in it, and make a mossy stone material for the ground." ## What is only on the website The connector focuses on generation and the most common follow up steps. For these, open the model in your [dashboard](https://www.3daistudio.com/Dashboard): - Advanced generation settings and all other 3D models on the platform - Texturing, the Texture Painter, rigging and animation, segmentation and remeshing with all options - 3D printing tools such as Mesh Repair, Add Base and Export for 3D Printing - Flow, the node based editor See the [rest of the documentation](https://docs.3daistudio.com/) for those tools. ## Frequently Asked Questions ### Which 3D model should I ask for? If you are unsure, do not specify one. Your assistant uses **Prism 3.1** by default, which gives the best overall quality. Ask for **P2** for low poly or quad meshes, **Hunyuan 3.1** for clean meshes at a flat price, and **Meshy 7.1** for very high resolution textures from an image. ### Can I use models I made on the website? Yes. Your assistant can browse your library and use any of your models or images as input, for example to convert, unwrap or make a low poly version. ### Why are there no prices in my assistant's replies? The connector asks assistants not to talk about credits unless you ask, to keep conversations focused. Ask "how much will this cost?" or "what's my balance?" at any time and it will tell you. --- # Images & Uploads Source: https://docs.3daistudio.com/mcp/images-and-uploads Many of the best results start from an image: a sketch, a product photo, concept art, or a reference you like. This page explains how to get your image to 3D AI Studio through your assistant. ## Why your assistant needs an upload step When you attach an image in a chat, your assistant can **see** it, but it cannot hand that file to other tools. 3D AI Studio therefore receives your image through a short upload step. It takes a few seconds and keeps your image at full quality. > [!NOTE] > **Just attach your image and say what you want.** For example: "turn this sketch into a 3D model". Your assistant knows it needs the upload step and starts it for you. ## Four ways to provide an image | Way | When to use it | How | | --- | --- | --- | | **Upload card in the chat** | Claude and other clients that show interactive cards | Your assistant opens a card. Drop your file in | | **Browser upload page** | Terminal and editor clients (Claude Code, Codex, Cursor, OpenCode, VS Code) | Your assistant gives you a link. Open it, drop your file in | | **An image link** | The image is already online at a public address | Paste the link (starting with `https://`) into the chat | | **An image from your library** | You generated or used the image before | Say which one, for example "use the wizard image I generated earlier" | ## Uploading with the card (Claude and similar clients) ### Ask for what you want Attach or describe your image and your goal, for example "turn this photo into a 3D model". ### Drop the file into the card Your assistant opens an **upload card** in the conversation. Drag your image onto it, or click to choose a file. ### Tell your assistant you are done Some assistants notice the upload automatically. If yours waits, just say **"done"** or **"uploaded"**. It then finds your upload and continues. ## Uploading with the browser page (terminal and editor clients) Terminal and editor clients cannot show cards, so the upload happens in your browser instead. ### Ask for what you want For example: "use this sketch to make a 3D model". If your assistant mentions an upload card that you cannot see, reply **"I don't see an upload card"**. ### Open the upload link Your assistant gives you a link to a 3D AI Studio upload page. Open it in your browser and drop your image in. ### Tell your assistant you are done Go back to your assistant and say **"done"**. It picks up your upload and continues. > [!WARNING] > Each upload link belongs to your account, works for a single upload, and expires after about 10 minutes. Do not share it. If a link has expired, ask your assistant for a new one. ## Supported files | | | | --- | --- | | Formats | PNG, JPEG, WebP | | Maximum file size | 20 MB | | Resolution | Large images are automatically scaled down to 2,500 pixels on the longest side, which is more than enough for every model | | How long uploads are kept | 24 hours, ready to be used in generations. Results you create from them are saved in your dashboard as usual | | Upload allowance | A generous number of uploads per hour, depending on your plan. If you reach it, the message tells you the limit; wait a bit and try again | ## Multiple images - **Image editing** can use several images at once: up to 10 with GPT Image 2 and up to 14 with Nano Banana. For example: "put the character from the first image into the scene from the second". - **Multi view 3D** uses 2 to 4 images of the same object (up to 8 with Hunyuan 3.1). Upload them in the order **front, left, back, right**, and say that they are views of one object. - If you upload several files, give each a clear name in the chat ("this one is the front view"). Your assistant can list your recent uploads and pick the right ones. ## Tips for images that become great 3D models - **One subject per image**, fully visible, not cut off at the edges. - **Clear lighting** and a reasonably sharp image. - **A simple background helps but is not required.** You do not need to remove the background before generating 3D. - **For characters**, a front facing pose with arms slightly away from the body works best. - **Start from an image you like.** If your source image is not quite right, ask your assistant to edit it first ("make the lighting even and the background plain"), then turn the edited image into 3D. See [Recommended Workflow](https://docs.3daistudio.com/3d-generation/recommended-workflow) for more on the image first approach. ## Frequently Asked Questions ### Why can't my assistant use the image I attached directly? Chat apps let the assistant look at attachments, but they do not pass the actual file on to connected tools. The upload step sends the original file to 3D AI Studio at full quality. ### My assistant keeps asking me to upload again. What should I do? Upload once, then say "done". If it still cannot find the image, ask it to "list my recent uploads" and tell it which one to use. ### Who can see my uploads? Only you. Uploads are tied to your account and are only used for the generations you ask for. See [Security & Privacy](https://docs.3daistudio.com/mcp/security-privacy). --- # Credits & Limits Source: https://docs.3daistudio.com/mcp/credits-and-limits The connector uses your normal 3D AI Studio credits. This page explains exactly when credits are spent, how to keep control of your spending, and which limits you might run into. ## How credits work through an assistant - **Same credits, same prices.** Every generation costs exactly what it costs on the website. There is no surcharge and no separate wallet. See [What You Can Do](https://docs.3daistudio.com/mcp/capabilities) for prices. - **Charged when a job starts.** Credits are deducted when 3D AI Studio accepts a generation, not when the assistant merely talks about it. - **Automatic refunds.** If a generation fails, the credits are refunded automatically, exactly as on the website. - **No double charges for repeats.** If your assistant accidentally sends the same request twice in quick succession, it runs once and you are charged once. - **Free actions.** Checking your balance, listing models, browsing your library, showing results, organizing projects and uploading images cost nothing. ## Personal or team credits When you connect an assistant, the sign in page asks **Use credits from**: - **Personal** (the default): generations use your own credits. - **A team** you belong to: generations use that team's credits. The choice is saved for that connection. Teams show up as a badge next to the connection in [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp). To switch, disconnect the assistant and connect it again, then pick the other option. If you leave a team, connections that used its credits stop working and show **Team access lost**. Disconnect and connect again to choose a new credit source. ## Staying in control of spending - **Your assistant only generates when you ask.** It does not start generations on its own. - **Ask before you spend.** "How much would that cost?" or "what's my balance?" are always free and always answered. - **Large requests ask first.** Unusually expensive requests need your confirmation before any credits are spent. Depending on your assistant, you see an approval prompt or the assistant asks you in the chat. Routine generations run straight away so the conversation does not stall. - **See what each assistant spent.** [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp) lists every connected assistant with when it was last used, how many actions it ran, and how many credits it has spent. - **Stop it at any time.** Disconnecting an assistant takes effect immediately. See [Managing Connections](https://docs.3daistudio.com/mcp/managing-connections). > [!NOTE] > To keep chats focused, assistants are asked not to bring up credit amounts unless you ask. If you want cost updates while you work, just say "tell me what each generation costs". ## Not enough credits If a request costs more than your balance, nothing is charged. Your assistant tells you the cost and your current balance, for example: > "This would cost 70 credits but the balance is only 40." Top up or change your plan on the [pricing page](https://www.3daistudio.com/Pricing), then ask your assistant to try again. ## Limits These limits are the same whether you use the website or an assistant. They exist to keep generation fast and reliable for everyone. | Limit | What happens | What to do | | --- | --- | --- | | **Generations per minute, per model** | Starting many generations with the same model in a short burst eventually returns "You've reached the limit of ... generations per minute for this model. Please wait ... seconds" | Nothing. Your assistant waits the number of seconds in the message and continues. If it gives up, ask it to try again | | **Uploads per hour** | Uploading many images in a short time returns "Upload limit reached (... per hour on the current plan)" | Wait a little, then upload again. Higher plans have a larger allowance | | **Images per request** | Up to 4 images per generation request, up to 10 (GPT Image 2) or 14 (Nano Banana) input images for edits | Split bigger jobs into several requests | | **Multi view input** | 2 to 4 views for most 3D models, up to 8 for Hunyuan 3.1 | Use the most informative views | > [!NOTE] > **Batching many models?** Tell your assistant something like "generate these 12 props one after another". It paces the requests and picks up where it left off if it hits a per minute limit. ## Plan requirement The connector is part of every **paid plan**. - **Free accounts** can add the connector, but the sign in page shows "A paid plan is required" instead of granting access. - **If your subscription ends**, your assistants stop working and show **Plan required** in Settings. Your models stay in your dashboard. After you upgrade again, the same connections work again; if one does not, connect it again. [See plans](https://www.3daistudio.com/Pricing) ## Frequently Asked Questions ### Does using an assistant cost more than the website? No. The prices are identical, and the connector itself is included in every paid plan. ### Can an assistant spend all my credits? It can only spend credits on generations you ask for, and unusually expensive requests need your confirmation. You can see each assistant's spending in Settings and disconnect it at any time. If you want extra safety, connect it to a team with a limited credit balance. ### Are failed generations charged? No. Failed generations are refunded automatically, the same as on the website. ### Why did my assistant pause before continuing a batch? It most likely hit the per minute limit for that model. It waits the time given in the message and then carries on. --- # Managing Connections Source: https://docs.3daistudio.com/mcp/managing-connections Every assistant you connect shows up in your 3D AI Studio settings. From there you can see what each one has done, fix connections that need attention, and disconnect any of them instantly. **Open it:** [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp) ## The AI Assistants (MCP) settings page The page has three parts: 1. **About the connector:** what assistants can do with 3D AI Studio, and an example prompt. 2. **Get connected:** the connector URL with a **Copy** button, and short setup steps for Claude, ChatGPT, Cursor and Claude Code, Codex, and OpenCode. 3. **Connected Apps:** every assistant currently connected to your account. On a free plan, the connector URL is shown with an upgrade button, because the connector is part of the paid plans. ## Connected Apps Each connected assistant is listed with: | Detail | Meaning | | --- | --- | | **Name** | The assistant that connected, for example Claude, ChatGPT, Cursor or Codex | | **Team badge** | Shown when the connection spends a team's credits instead of your personal credits | | **Status badge** | Shown only when the connection needs attention (see below) | | **Last used** | "Used today", "Used 3 days ago", or "Never used" | | **Tool calls** | How many actions the assistant has run through 3D AI Studio | | **Credits spent** | How many credits this connection has spent in total | | **Disconnect** | Removes the assistant's access immediately | Each connection is one app on one account. Desktop and terminal tools sometimes register separately per installation, so the same tool can appear more than once. That is normal; disconnect the ones you no longer use. ## Status badges and how to fix them Most connections never show a badge. When one appears, the settings page also shows a short hint with the fix, and for terminal tools a command you can copy. | Badge | What it means | How to fix it | | --- | --- | --- | | **Sign in again** | The assistant's sign in has ended (for example after disconnecting, a long break, or a sign out everywhere) | **Chat apps:** open the assistant's connector settings and connect 3D AI Studio again. **Codex:** run `codex mcp login 3daistudio`. **OpenCode:** run `opencode mcp auth 3daistudio`. **Claude Code:** run `/mcp` inside Claude Code, pick the 3D AI Studio server and choose Authenticate | | **Restart needed** | One of the assistant's windows or sessions is still using an old sign in. The connection itself is fine | Close all running sessions of that assistant and start a new one (for chat apps, restart the app) | | **Plan required** | Your account no longer has an active paid plan | Upgrade with the **View plans** button. The connection works again afterwards | | **Team access lost** | The connection spends credits from a team you are no longer part of | Disconnect it, then connect the assistant again and pick a new credit source | If you named the server something other than `3daistudio` in a terminal tool, use your own name in the commands. ## Disconnecting an assistant Click **Disconnect** next to the assistant in [Connected Apps](https://www.3daistudio.com/Dashboard?settings=mcp). - Access ends **immediately**. Every credential that assistant held is deleted, so its next request fails. - Your models, images and projects are **not** affected. Everything the assistant created stays in your dashboard. - To use the assistant again later, simply connect it again. You can also remove the server from the assistant's own settings (for example `claude mcp remove 3daistudio` in Claude Code). That only removes it from the assistant. To be sure the access is gone, also disconnect it in 3D AI Studio. > [!NOTE] > **Sign out everywhere also disconnects assistants.** Using **Sign out of all devices** in Settings → Security ends every login session and disconnects every connected assistant at once. Use it if you think someone else has access to your account. ## Switching between personal and team credits The credit source is chosen when you connect. To change it: 1. Disconnect the assistant in Connected Apps. 2. Connect it again from the assistant. 3. On the sign in page, choose the other option under **Use credits from**. ## Frequently Asked Questions ### How do I see which assistants have access to my account? Open [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp) and look at Connected Apps. Every assistant with access is listed there. ### Does disconnecting delete my models? No. Disconnecting only removes the assistant's access. All generations stay in your dashboard. ### An assistant I don't recognize is listed. What should I do? Click **Disconnect** right away, then use **Sign out of all devices** in Settings → Security and change your password if you use one. Contact [support@3daistudio.com](mailto:support@3daistudio.com) if you see activity you did not start. --- # MCP Troubleshooting Source: https://docs.3daistudio.com/mcp/troubleshooting Find the message or symptom you see, then follow the fix. The messages are quoted the way your assistant receives them, so you can also search this page for the exact words. > [!NOTE] > **First check:** open [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp). If your assistant shows a badge (Sign in again, Restart needed, Plan required, Team access lost), the hint below it tells you the fix. See [Managing Connections](https://docs.3daistudio.com/mcp/managing-connections). ## Quick reference | You see | Jump to | | --- | --- | | The connector cannot be added, or the server shows an error | [Connection problems](https://docs.3daistudio.com/mcp/troubleshooting#connection-problems) | | "A paid plan is required" or "requires an active paid subscription" | [Plan required](https://docs.3daistudio.com/mcp/troubleshooting#plan-required) | | Asked to sign in again, or the server needs authentication | [Signed out](https://docs.3daistudio.com/mcp/troubleshooting#signed-out-or-needs-authentication) | | "Restart needed" in Settings | [Restart needed](https://docs.3daistudio.com/mcp/troubleshooting#restart-needed) | | "You've reached the limit of ... generations per minute" | [Rate limits](https://docs.3daistudio.com/mcp/troubleshooting#rate-limits) | | "Not enough credits" | [Not enough credits](https://docs.3daistudio.com/mcp/troubleshooting#not-enough-credits) | | An image could not be used | [Image problems](https://docs.3daistudio.com/mcp/troubleshooting#image-problems) | | "generation failed" | [Failed generations](https://docs.3daistudio.com/mcp/troubleshooting#failed-generations) | | The assistant does not use 3D AI Studio | [Assistant ignores the tools](https://docs.3daistudio.com/mcp/troubleshooting#the-assistant-does-not-use-3d-ai-studio) | | No 3D viewer in the chat | [Result display](https://docs.3daistudio.com/mcp/troubleshooting#no-3d-viewer-in-the-chat) | ## Connection problems **Symptoms:** adding the connector fails, the server shows red or "error" in your client, a "404" or "not found", or the client says it is not an MCP server. **Fixes:** 1. **Check the URL.** It must be exactly `https://mcp.3daistudio.com/mcp`. - Not `https://www.3daistudio.com` (that is the website, not the connector). - Not `https://mcp.3daistudio.com` without `/mcp` at the end. - Not an address ending in `/sse` from an older guide. 2. **Check the connection type.** It must be a **remote** server using **Streamable HTTP** (often just called "http" or "remote"), with **OAuth** sign in. Remove any API key, token or custom header you added. 3. **Update your client.** Remote servers with sign in need a recent version of Claude Code, Codex, Cursor, VS Code or OpenCode. 4. **Restart the client** after changing its MCP configuration. 5. **Ask your assistant.** AI apps change how they add MCP servers often, so our guides can lag behind. Ask your assistant: "How do I add a remote MCP server in this app right now? Use https://docs.3daistudio.com/mcp-setup.md." Coding agents can usually fix the setup themselves. **Error on the sign in page ("Connection error")**: close the window and start the connection again from your assistant. If it keeps happening, remove the connector from your assistant, add it again, and make sure the URL is correct. ## Plan required **Messages:** > "A paid plan is required. Connecting AI assistants like Claude, ChatGPT, or Cursor to your 3D AI Studio account is available on paid plans." > "The 3D AI Studio MCP connector requires an active paid subscription. Upgrade at https://www.3daistudio.com/pricing to keep using it." **Fix:** the connector is part of the paid plans. [Upgrade your plan](https://www.3daistudio.com/Pricing), then connect again (or just retry, if the assistant was already connected). Your existing models are never affected. If you are part of a team, check that you chose the right credit source when connecting. See [Credits & Limits](https://docs.3daistudio.com/mcp/credits-and-limits#personal-or-team-credits). ## Signed out or needs authentication **Symptoms:** the assistant asks you to connect again, the server shows "needs authentication" or "login required", or Settings shows **Sign in again**. **Fix by assistant:** | Assistant | What to do | | --- | --- | | Claude | Settings → Connectors → 3D AI Studio → Connect | | ChatGPT | Settings → Connectors → 3D AI Studio → connect again | | Claude Code | Run `/mcp`, select `3daistudio`, choose **Authenticate** | | Codex | Run `codex mcp login 3daistudio` | | OpenCode | Run `opencode mcp auth 3daistudio` | | Cursor | Cursor Settings → MCP → use the login or connect action next to `3daistudio` | | VS Code | Run **MCP: List Servers**, restart `3daistudio` and accept the sign in prompt | **Why it happens:** you disconnected the assistant, used **Sign out of all devices**, or the assistant was not used for a long time. Signing in again restores everything. **Keeps happening with Codex or another terminal tool?** Close **all** running sessions of the tool, update it to the latest version, then sign in once more. Several old sessions running at the same time are the most common cause. ## Restart needed **Symptom:** Settings shows **Restart needed** next to an assistant. **What it means:** one window or session of that assistant is still using an old sign in. The connection itself is fine, and other sessions keep working. **Fix:** close all running sessions or windows of the assistant and start a new one. For chat apps, restart the app. ## Rate limits **Message:** > "You've reached the limit of ... generations per minute for this model. Please wait ... seconds before starting another generation." **What it means:** many generations with the same model were started in a short time. This limit is the same as on the website. **Fix:** usually nothing. Your assistant waits the number of seconds in the message and continues. If it stops instead, say "please continue" or "try again in a minute". For big batches, ask it to generate the items one after another. **Upload limit:** > "Upload limit reached (... per hour on the current plan). Please try again later." Wait a while before uploading more images. Higher plans have a larger hourly allowance. ## Not enough credits **Messages:** > "Not enough credits for this ... generation. Current balance: ..." > "This would cost ... credits but the balance is only ..." **Fix:** nothing was charged. Top up or upgrade on the [pricing page](https://www.3daistudio.com/Pricing), or choose a cheaper option (for example Prism 3.1 instead of P2, or a lower image quality). If you connected with team credits, the team's balance is what counts. ## Image problems | Message | Fix | | --- | --- | | "Images must be PNG, JPEG, or WebP." | Convert the image to PNG, JPEG or WebP and upload it again | | "Image exceeds the 20MB limit." | Use a smaller or more compressed image. Large images are scaled down automatically anyway, so you lose nothing | | "Image URLs must point to a public host." or "Could not resolve image host" | The link is not publicly reachable (for example a local or intranet address). Upload the file instead. See [Images & Uploads](https://docs.3daistudio.com/mcp/images-and-uploads) | | "Malformed upload reference" | Ask your assistant to "list my recent uploads" and use the reference exactly as listed | | "Generation ... has no stored image to reuse (it may still be processing)." | Wait until the source image has finished, then try again | | The upload link says it has expired | Upload links work once and expire after about 10 minutes. Ask your assistant for a new one | | You cannot see an upload card | Tell your assistant "I don't see an upload card". It gives you a browser link instead | | The assistant cannot find your upload | Say "list my recent uploads" and tell it which one to use | ## Failed generations **Messages:** > "... generation failed: ... Credits were refunded automatically." > "An identical recent request already failed: ... Credits were refunded. Vary the arguments (e.g. the prompt or a seed) to retry." **Fix:** your credits are already back. Try again with a slightly different prompt or image, or ask for a different model (for example "try it with Hunyuan 3.1"). If the image itself is the problem, see the image tips in [Images & Uploads](https://docs.3daistudio.com/mcp/images-and-uploads#tips-for-images-that-become-great-3d-models). **"An identical generation is already in progress"**: the same request was sent twice. Nothing extra is charged. Wait for the first one to finish; your assistant can check it with "how is my generation doing?". ## The assistant does not use 3D AI Studio **Symptoms:** the assistant writes code for a 3D shape, describes a model in words, or says it cannot create 3D models. **Fixes:** 1. **Enable the connector in the chat.** In Claude and ChatGPT, turn on 3D AI Studio in the tools menu of the message box. In Cursor and VS Code, use agent mode and make sure the 3daistudio tools are enabled. 2. **Ask explicitly:** "Use 3D AI Studio to make a 3D model of ...". 3. **Check the connection:** ask "what's my 3D AI Studio credit balance?". If that fails, go through [Connection problems](https://docs.3daistudio.com/mcp/troubleshooting#connection-problems) and [Signed out](https://docs.3daistudio.com/mcp/troubleshooting#signed-out-or-needs-authentication). 4. **Start a new chat or session** after adding the connector. Many clients only load tools at the start. ## No 3D viewer in the chat The interactive 3D viewer appears in assistants that support interactive cards, such as Claude. Terminal and editor clients (Claude Code, Codex, Cursor, VS Code, OpenCode) show a **link** instead, and some chat apps show a simple card or link. In every case the model is saved in your [dashboard](https://www.3daistudio.com/Dashboard), where you can view it in 3D and download it in any format. ## Results are missing from the dashboard Every generation the assistant starts is saved to the dashboard of the account (and team, if you chose one) it is connected to. - **Check the team.** If you connected with team credits, the results are in that team's library. - **Check the account.** Make sure you signed in with the same 3D AI Studio account on the website as during the connection. - **Wait for it to finish.** 3D models take 1 to 5 minutes. They appear in the dashboard right away and complete there. ## Still stuck? Email [support@3daistudio.com](mailto:support@3daistudio.com) with: - the assistant you use and its version (for example "Codex CLI 0.145" or "Claude desktop"), - what you asked for and the exact message you saw, - roughly when it happened. Never send your password, tokens or the contents of configuration files with credentials. We never need them. For general (non MCP) problems, see the main [Troubleshooting Guide](https://docs.3daistudio.com/troubleshooting). --- # Security & Privacy Source: https://docs.3daistudio.com/mcp/security-privacy Connecting an assistant gives it limited, revocable access to your 3D AI Studio account. This page explains in plain language what that access includes, what it does not, and how to stay in control. ## How the connection is secured - **You sign in on 3D AI Studio, not in the assistant.** The sign in page belongs to 3D AI Studio. Your assistant never sees your password, passkey or one time codes. - **No API key.** Access uses the industry standard OAuth sign in. Nothing secret is created for you to copy, paste or store in a configuration file, so there is nothing that can leak from a file or a repository. - **You approve every connection.** Each assistant shows up on its own sign in page with its name, and access is only granted after you click **Authorize**. - **Encrypted connection.** All traffic between your assistant and 3D AI Studio uses HTTPS. - **Revocable at any time.** You can disconnect any assistant instantly. See [Managing Connections](https://docs.3daistudio.com/mcp/managing-connections). > [!WARNING] > **Only authorize connections you started yourself.** The sign in page shows the name of the app asking for access and, where available, its web address. If you did not just connect an assistant, or the name or address looks wrong, click **Cancel**. ## The permissions an assistant gets The sign in page lists these under "This app will be able to": | Permission | What it allows | | --- | --- | | **Generate 3D models using your credits** | Start 3D generations and 3D processing (low poly, UV unwrap, format conversion) | | **Generate and edit images using your credits** | Generate and edit images, create materials, remove backgrounds, and receive the images you upload | | **View your generations and projects** | Browse your library, check the status of generations, and show results | | **Create projects and organize your generations** | Create dashboard projects and move generations into them | | **View your credit balance** | Read your balance and the model price list | Spending always comes from the credit source you chose when connecting: your **personal** credits or **one team** you belong to. ## What an assistant can never do - See or change your password, passkeys, email address or other account settings. - See your payment details, or change your plan, billing or subscription. - Delete your models, images or projects. - Make your models public or share them. - Invite or remove team members, or change team settings. - Spend credits from a credit source you did not choose. ## Safeguards for your credits - Assistants only generate when you ask. Every generation costs the same as on the website, and failed generations are refunded automatically. - Unusually expensive requests need your confirmation before any credits are spent. - [Settings → AI Assistants (MCP)](https://www.3daistudio.com/Dashboard?settings=mcp) shows how many actions each assistant has run and how many credits it has spent. - For shared or automated agents, connect them to a **team** with a limited credit balance, so they can never spend more than you put there. More details are on [Credits & Limits](https://docs.3daistudio.com/mcp/credits-and-limits). ## Your data - **Your generations stay private.** Models and images created through an assistant are saved to your dashboard with the same privacy as everything you create on the website. They are not published, and they are not used for training. - **Uploads are only for you.** Images you upload through the connector belong to your account, are only used for the generations you request, and are removed after 24 hours. Results you create from them stay in your dashboard. - **What your assistant's provider sees.** The assistant you use (for example Claude by Anthropic or ChatGPT by OpenAI) sees your conversation, including what 3D AI Studio sends back to it, such as the status of a generation, result links and previews. That part is covered by your assistant provider's own privacy terms. ## Good habits - **Review your connections now and then.** Disconnect assistants you no longer use. - **Never paste credentials into a chat or configuration file.** The connector never needs a password, token or API key. If an agent or a setup guide asks for one, it is not following the official setup. - **Use the official address only.** The connector URL is `https://mcp.3daistudio.com/mcp`. Do not connect to look alike addresses. - **If something looks wrong,** disconnect the assistant, use **Sign out of all devices** in Settings → Security (this also disconnects every assistant), and contact us. ## Reporting a security issue If you believe you found a security problem with the connector, email [support@3daistudio.com](mailto:support@3daistudio.com) with the subject "Security". Please do not post details publicly before we have had a chance to look at it. ## Frequently Asked Questions ### Is it safe to connect my account to an AI assistant? Yes. The assistant gets only the permissions listed above, never your password, and you can disconnect it instantly. The main thing to watch is spending, which you can control with the safeguards above. ### Can the assistant see my other files or projects? It can see your 3D AI Studio library and projects so it can work with your models. It cannot access anything outside your 3D AI Studio account through the connector. ### Can someone else use my connection? Only someone with access to your assistant account or your computer's assistant session. Protect those like any other account. If you suspect misuse, disconnect the assistant and sign out of all devices. ### Does 3D AI Studio see my chat with the assistant? No. 3D AI Studio only receives the requests the assistant sends to the connector, such as "generate a 3D model with this prompt". It does not receive the rest of your conversation. --- # Integrations Source: https://docs.3daistudio.com/integrations 3D AI Studio models work with all major 3D software. Export as GLB, FBX, or OBJ and import into your tool of choice. See [Export Formats](https://docs.3daistudio.com/export-formats) for details on each format. ## Supported Software | Software | Best Format | Use Case | |----------|-------------|----------| | [Blender](https://docs.3daistudio.com/integrations/blender) | GLB or FBX | Import, edit materials, rig, animate | | [Unity](https://docs.3daistudio.com/integrations/unity) | FBX or GLB | Import for game projects | | [Unreal Engine](https://docs.3daistudio.com/integrations/unreal) | FBX | Import with materials for UE4/UE5 | | [ZBrush](https://docs.3daistudio.com/integrations/zbrush) | OBJ | Import for sculpting and detailing | | [Maya](https://docs.3daistudio.com/integrations/maya) | FBX or OBJ | Production pipelines, animation | | [Godot](https://docs.3daistudio.com/integrations/godot) | GLB | Import for game dev in Godot 4 | ## Video: AI 3D Models in Game Engines [Video: AI 3D Models in Game Engines](https://www.youtube.com/watch?v=L5emNlNYVfY) ## Quick Start 1. Generate your model using [Image to 3D](https://www.3daistudio.com/ImageTo3D) or [Text to 3D](https://www.3daistudio.com/TextTo3D) 2. Download in the right format for your software 3. Import into your tool using the guide above 4. Adjust materials, scale, and lighting as needed For polygon reduction before import, use the [Remesh tool](https://www.3daistudio.com/Tools/Remesh). For rigged characters, check the [Rigging tool](https://www.3daistudio.com/Tools/RiggingTool). ## AI Assistants (MCP) You can also generate models without leaving your AI assistant. The [MCP connector](https://docs.3daistudio.com/mcp/overview) works with Claude, ChatGPT, Cursor, Claude Code, Codex, VS Code (Copilot) and other MCP clients, and can export directly to FBX, OBJ, STL, USDZ or 3MF for the software above. Create a model with the [AI 3D model generator](https://www.3daistudio.com) and import it into your favorite software. ## Frequently Asked Questions ### What is the best file format for importing AI-generated 3D models into a game engine? FBX is the safest choice for Unity and Unreal Engine because both engines handle FBX materials, rigging, and animations natively. 3D AI Studio lets you download as FBX directly, so you can drag the file into your project without any conversion step. For Godot, use GLB instead since Godot 4 has built-in glTF support. ### Do AI-generated 3D models work in professional 3D software like Maya and ZBrush? Yes. Models from 3D AI Studio export as standard GLB, FBX, and OBJ files that open in any professional 3D application. Maya reads FBX with full material data, ZBrush imports OBJ for sculpting, and Blender handles all three formats. The files follow the same standards as any other 3D asset. ### Can I edit an AI-generated model after importing it into my 3D software? Absolutely. Once you import a model from 3D AI Studio, it behaves like any other mesh in your software. You can edit geometry, swap textures, add rigging, animate, or combine it with other assets. Use the Remesh tool before downloading if you need cleaner topology for editing. ### Do I need to install anything to use AI-generated 3D models? No special plugins are required for most software. 3D AI Studio exports in standard formats that Blender, Unity, Unreal Engine, Maya, and ZBrush all support out of the box. The only exception is if you want to use GLB in Unity, which needs a free glTF importer package. --- # Unity Integration Source: https://docs.3daistudio.com/integrations/unity To import a 3D AI Studio model into Unity, download it as **FBX** and drag the file into your project's **Assets** folder. Unity imports the mesh, materials and textures automatically and creates a prefab you can drop into any scene. | | | | --- | --- | | **Recommended format** | FBX (GLB needs the GLTFast or UnityGLTF package) | | **Import** | Drag the file into the Assets folder | | **Shaders** | Standard, URP and HDRP pick up the textures | | **Characters** | Rigged FBX, Animation Type Humanoid | | **Mobile budget** | Under 10,000 triangles per model | ## How do I import a 3D model into Unity? Drag the FBX file into the **Assets** folder. Unity imports it and creates a prefab, which you then drag into your scene. 1. Download the model as FBX from your [Dashboard](https://www.3daistudio.com/Dashboard) 2. Drag the .fbx file into your Unity **Assets** folder (or a subfolder) 3. Unity imports the model and creates a prefab 4. Drag the prefab into your scene ## Should I use FBX or GLB in Unity? Use FBX: Unity supports it natively with materials, rigging and animation. GLB works too, but only after installing a glTF package. Only have a GLB from somewhere else? Convert it with the free [GLB to FBX converter](https://www.3daistudio.com/convert/glb-to-fbx). | | FBX | GLB | | --- | --- | --- | | Unity support | Native | Needs **GLTFast** (recommended, via Package Manager) or **UnityGLTF** | | Materials | Picked up by Standard, URP and HDRP shaders | Handled by the package | | Rigging and animation | Full support | Supported by the package | After installing a glTF package, drag .glb files into Assets like any other model. ## Why do my materials look wrong in Unity? Usually the textures were not assigned to the shader slots. Select the model, open the **Materials** tab in the import settings, make sure **Use Embedded Materials** is selected, and reassign the albedo, normal and roughness textures to the URP or Standard shader slots if needed. FBX materials from 3D AI Studio come with albedo, normal and roughness textures that Unity's Standard and URP Lit shaders read automatically in most cases. ## How do I set up a rigged character in Unity? Import the rigged FBX, open the **Rig** tab, set **Animation Type** to **Humanoid** and click **Apply**. Unity maps the bones to its Humanoid rig, so Mixamo animations and your own clips work through an Animator Controller. 1. Import the FBX from the [Rigging tool](https://docs.3daistudio.com/processing/rigging) 2. Select the model in the Project panel 3. Open the **Rig** tab in the Inspector 4. Set **Animation Type** to **Humanoid** for characters, or **Generic** for other creatures 5. Click **Apply** ## How many polygons should a model have for Unity? Keep models under 10,000 triangles for mobile games. On PC and console, hero assets can use 50,000-100,000 triangles. Set the exact target with the [Remesh tool](https://docs.3daistudio.com/processing/remesh) before downloading. Further optimization in Unity: - Set up **LOD Groups** for models seen at different distances - Enable **mesh compression** in the import settings for a smaller build - Reduce polygons before import to save draw calls and memory ## Common workflows - **Props and environments:** import a static FBX, add colliders and place it in the scene - **Characters:** import a rigged FBX, set it up as Humanoid and attach an Animator - **Product visualization:** import a detailed model and set up camera and lighting Don't have a model yet? Generate one with [Image to 3D](https://www.3daistudio.com/ImageTo3D) or [Text to 3D](https://www.3daistudio.com/TextTo3D) and import it into Unity in minutes. Format details are in [Export Formats](https://docs.3daistudio.com/export-formats). ## Frequently Asked Questions ### Do the models work with URP and HDRP? Yes. The FBX materials include albedo, normal and roughness textures that the URP Lit shader picks up automatically. If a material looks wrong, select "Use Embedded Materials" in the import settings or reassign the textures to the shader slots. ### Can an AI agent work inside the Unity Editor? Yes, with a Unity MCP server. It lets Claude Code, Cursor or Codex create GameObjects, edit C# and run tests, while the [3D AI Studio MCP connector](https://docs.3daistudio.com/mcp/overview) generates the models. Our [Unity MCP guide](https://www.3daistudio.com/3d-generator-ai-comparison-alternatives-guide/unity-mcp) compares Unity's own server with the community options. --- # Unreal Engine Integration Source: https://docs.3daistudio.com/integrations/unreal To import a 3D AI Studio model into Unreal Engine, download it as **FBX** and drag it into the **Content Browser**, then click **Import All**. Unreal creates the mesh, materials and textures automatically, and in UE5 you can enable Nanite to use high-poly models without optimizing them first. | | | | --- | --- | | **Recommended format** | FBX | | **Import** | Drag into the Content Browser, then Import All | | **Materials** | Material instances created automatically from PBR textures | | **High-poly static meshes** | Nanite (UE5) | | **Characters** | Imported as Skeletal Mesh | ## How do I import a 3D model into Unreal Engine? Drag the FBX into the **Content Browser**, keep mesh, texture and material import enabled, and click **Import All**. The model is then ready to place in a level. 1. Download the model as FBX from your [Dashboard](https://www.3daistudio.com/Dashboard) 2. Open your Unreal project 3. Drag the .fbx file into the **Content Browser** 4. In the FBX Import Options dialog, enable **Import Mesh**, **Import Textures** and **Import Materials** 5. Click **Import All** FBX is the recommended format because Unreal imports its materials, textures, skeletal meshes and animations natively. OBJ also imports but loses material data. ## Why are my textures missing in Unreal? A texture slot was not connected on import. Open the imported material and drag each texture from the Content Browser into its slot. | Material slot | Texture | | --- | --- | | Base Color | Albedo / diffuse | | Normal | Normal map | | Roughness | Roughness map | | Metallic | Metallic map | In most cases Unreal connects these automatically, and PBR textures from the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) map well to Unreal's material system. ## Can I use Nanite with AI-generated models? Yes. Nanite handles high-poly static meshes with millions of triangles and builds LODs automatically. After importing, right-click the static mesh in the Content Browser and choose **Nanite > Enable**. Nanite works on static meshes only. For skeletal meshes and UE4 projects, reduce the polygon count with the [Remesh tool](https://docs.3daistudio.com/processing/remesh) before importing. ## How do I set up a rigged character in Unreal? Import the rigged FBX from the [Rigging tool](https://docs.3daistudio.com/processing/rigging). Unreal detects it as a **Skeletal Mesh**, and you can retarget animations from the Unreal Mannequin or use your own. 1. Import the FBX. Unreal detects the Skeletal Mesh automatically 2. Choose an existing skeleton or let Unreal create one 3. Retarget animations from the Unreal Mannequin, or apply your own 4. Set up an Animation Blueprint for gameplay ## How do I keep performance good in Unreal? - Enable **Nanite** for high-poly static meshes in UE5 - **Generate lightmap UVs** on import for baked lighting - **Auto-generate collision** for interactive objects - Use **LOD groups** in open-world scenes - Reduce polygons with the [Remesh tool](https://docs.3daistudio.com/processing/remesh) for skeletal meshes and UE4 ## Common workflows - **Game assets and environments:** import static meshes, add collision, place them in levels - **Architectural visualization:** import detailed models and light them with Lumen - **Virtual production:** combine Lumen and a virtual camera for real-time cinematics Don't have a model yet? Generate one with [Image to 3D](https://www.3daistudio.com/ImageTo3D) or [Text to 3D](https://www.3daistudio.com/TextTo3D) and import it into Unreal Engine in minutes. Format details are in [Export Formats](https://docs.3daistudio.com/export-formats). --- # Godot Integration Source: https://docs.3daistudio.com/integrations/godot To import a 3D AI Studio model into Godot 4, download it as **GLB** and drag the file into the **FileSystem** dock. Godot imports it as a scene with PBR materials and textures, ready to drag into your level. | | | | --- | --- | | **Recommended format** | GLB (native glTF 2.0 support) | | **Import** | Drag into the FileSystem dock | | **Materials** | Albedo, normal, roughness and metallic read automatically | | **Collision** | Mesh > Create Trimesh Static Body or Simplified Convex Shape | | **Mobile budget** | Under 10,000 triangles per model | ## How do I import a 3D model into Godot? Drag the GLB into the **FileSystem** dock. Godot imports it as a scene, which you then drag into the 3D viewport or scene tree. 1. Download the model as GLB from your [Dashboard](https://www.3daistudio.com/Dashboard) 2. Drag the .glb file into the **FileSystem** dock 3. Godot imports it as a scene 4. Drag the scene into your 3D viewport or scene tree ## Should I use GLB or FBX in Godot? Use GLB. Godot 4 supports glTF 2.0 natively, and the GLB keeps geometry, PBR materials and textures in one file. FBX also works but may need the FBX2glTF plugin depending on your setup. Simpler: turn the FBX into a GLB with the free [FBX to GLB converter](https://www.3daistudio.com/convert/fbx-to-glb). ## How do I edit the materials in Godot? Godot reads the PBR textures from the GLB automatically. To change them, select the **MeshInstance3D** node and edit or override the material in the Inspector. 1. Instance the scene in your level 2. Select the **MeshInstance3D** node 3. Expand the **Material** section in the Inspector 4. Edit the material, or override it with a new **StandardMaterial3D** ## How do I add collision to a model in Godot? Select the **MeshInstance3D** node and use the **Mesh** menu: **Create Trimesh Static Body** for complex shapes or **Create Simplified Convex Shape** for faster physics. Both add a StaticBody3D with a CollisionShape3D as children. For characters and moving objects, use **CharacterBody3D** or **RigidBody3D** instead of a static body. ## How do I keep performance good in Godot? Keep mobile models under 10,000 triangles and reduce high-poly models with the [Remesh tool](https://docs.3daistudio.com/processing/remesh) before downloading. - Use **MultiMeshInstance3D** to render many copies of the same model, such as trees or rocks - Check the scale after import and adjust the node's **Scale** to match your game's units - GLB keeps imports clean: one file, no missing textures ## Common workflows - **Props and game assets:** import the GLB, add a StaticBody3D with a collision shape, place it in the level - **Environments:** import several models, arrange them and add lighting - **Characters:** import a rigged GLB from the [Rigging tool](https://docs.3daistudio.com/processing/rigging) and play animations with an AnimationPlayer Don't have a model yet? Generate one with [Image to 3D](https://www.3daistudio.com/ImageTo3D) or [Text to 3D](https://www.3daistudio.com/TextTo3D) and import it into Godot in minutes. Format details are in [Export Formats](https://docs.3daistudio.com/export-formats). ## Frequently Asked Questions ### Can an AI agent work inside the Godot editor? Yes, with a Godot MCP server. It lets Claude Code, Cursor or Codex build scenes, edit GDScript and run your game, while the [3D AI Studio MCP connector](https://docs.3daistudio.com/mcp/overview) generates the models. Our [Godot MCP guide](https://www.3daistudio.com/3d-generator-ai-comparison-alternatives-guide/godot-mcp) compares the main servers. --- # Blender Integration Source: https://docs.3daistudio.com/integrations/blender To import a 3D AI Studio model into Blender, download it as **GLB** and use **File > Import > glTF 2.0 (.glb/.gltf)**. The model arrives with its textures and PBR materials already connected, ready for Cycles and Eevee. | | | | --- | --- | | **Recommended format** | GLB (FBX for rigged or animated models) | | **Import menu** | File > Import > glTF 2.0 (.glb/.gltf) | | **Materials** | Imported as a Principled BSDF with textures connected | | **Renderers** | Cycles and Eevee, no manual setup | ## How do I import a GLB into Blender? Go to **File > Import > glTF 2.0 (.glb/.gltf)**, select the .glb file and click **Import glTF 2.0**. Then switch to Material Preview to see the textures. 1. Download the model as GLB from your [Dashboard](https://www.3daistudio.com/Dashboard) 2. In Blender, open **File > Import > glTF 2.0 (.glb/.gltf)** 3. Select the .glb file 4. Click **Import glTF 2.0** 5. Press **Z** and choose **Material Preview** to see the textures ## How do I import an FBX into Blender? Go to **File > Import > FBX (.fbx)**, select the file and click **Import FBX**. Use FBX mainly for rigged models from the [Rigging tool](https://docs.3daistudio.com/processing/rigging), because it carries bones and animation data. 1. Open **File > Import > FBX (.fbx)** 2. Select the .fbx file 3. Click **Import FBX** ## Should I use GLB or FBX for Blender? Use GLB for most models: geometry, textures and materials come in one file, so nothing goes missing. Use FBX when the model is rigged or animated. | | GLB | FBX | | --- | --- | --- | | Textures | Embedded in the file | Included | | Materials | Principled BSDF connected automatically | Imported, may need checks | | Rigging and animation | Supported | The usual choice for rigs | | Best for | Static models, props, renders | Characters from the Rigging tool | ## How do I fix scale, missing textures or dark faces after import? Most import issues come down to scale, viewport mode or normals, and each has a one-step fix. | Problem | Fix | | --- | --- | | Model is huge or tiny | Select it, press **S** to scale, then **Ctrl+A > Scale** to apply | | No textures visible | Press **Z** and switch to **Material Preview** or Rendered view | | Faces look dark or inverted | In Edit Mode select all, then **Mesh > Normals > Recalculate Outside** | | Transforms cause problems later | Apply them with **Ctrl+A > All Transforms** before rigging or exporting | | Scene is too heavy | Reduce polygons with the [Remesh tool](https://docs.3daistudio.com/processing/remesh) before downloading | 3D AI Studio models use metric units, while Blender scenes sometimes use a different unit scale, which is why size differences happen. ## What can I do with the model in Blender? ### Edit materials Open the **Shader Editor**. GLB imports come with a Principled BSDF node connected to the textures, so you can change roughness or metallic values or swap textures directly. PBR materials from the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) carry over automatically. ### Rig and animate Import a pre-rigged FBX from the [Rigging tool](https://docs.3daistudio.com/processing/rigging), which arrives with bones ready, or add your own armature with Rigify or manual rigging. Apply all transforms before binding the mesh to bones. ### Render Render in **Cycles** or **Eevee**. The PBR maps (albedo, roughness, metallic, normal) map to Blender's shader system, so renders look correct without manual setup. ## How do I export from Blender to a game engine? Go to **File > Export** and choose **FBX** for Unity and Unreal Engine or **GLB** for Godot and the web. Apply all transforms first so the model keeps its size and orientation. Engine-specific steps are in the [Unity](https://docs.3daistudio.com/integrations/unity), [Unreal Engine](https://docs.3daistudio.com/integrations/unreal) and [Godot](https://docs.3daistudio.com/integrations/godot) guides. Don't have a model yet? Generate one with [Image to 3D](https://www.3daistudio.com/ImageTo3D) or [Text to 3D](https://www.3daistudio.com/TextTo3D) and import it into Blender in minutes. ## Frequently Asked Questions ### Can Claude work inside Blender? Yes. Blender Lab publishes an official Blender connector for Claude Desktop. Pair it with the [3D AI Studio MCP connector](https://docs.3daistudio.com/mcp/overview) and Claude can generate a model and arrange the scene around it. Our [Blender MCP guide](https://www.3daistudio.com/3d-generator-ai-comparison-alternatives-guide/blender-mcp) covers setup and what it can and cannot do. --- # Maya Integration Source: https://docs.3daistudio.com/integrations/maya ## Recommended Format - **FBX** - full material, hierarchy, and animation support - **OBJ** - geometry only, no materials or animation data Use FBX when you need materials to carry over. Use OBJ if you only need the mesh. ## Import Steps 1. Open Maya 2. Go to **File > Import** 3. Set the file type filter to FBX or OBJ 4. Select your downloaded file 5. Click **Import** The model appears in your viewport with materials attached (FBX) or as plain geometry (OBJ). ## Material Setup ### FBX Maya's Arnold shader (aiStandardSurface) reads FBX materials automatically. Open the Hypershade to see the imported materials and their texture connections. ### OBJ OBJ files import geometry only. You need to create materials manually: 1. Open the Hypershade (Windows > Rendering Editors > Hypershade) 2. Create a new aiStandardSurface material 3. Connect your texture files to Base Color, Normal, Roughness, and Metalness 4. Assign the material to your mesh PBR textures from the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) work well with Arnold's shader inputs. ## Common Workflows - **Character rigging and animation** - import FBX, build joint skeleton, bind skin, animate - **VFX and film production** - use as base geometry, add subdivision surfaces, render with Arnold - **Product visualization** - import model, set up studio lighting, render turntables ## Tips - FBX preserves much more data than OBJ (materials, hierarchy, blend shapes, animations) - prefer it when possible - Check your scene units (Window > Settings > Preferences > Settings > Working Units) before import to avoid scale issues - Apply Freeze Transformations after scaling to keep clean transform values - For characters, the [Rigging tool](https://docs.3daistudio.com/processing/rigging) in 3D AI Studio gives you a pre-rigged FBX that imports directly into Maya with bones ready ## Useful Links - [Remesh tool](https://docs.3daistudio.com/processing/remesh) - reduce polygon count - [Export Formats](https://docs.3daistudio.com/export-formats) - format comparison - [Rigging tool](https://docs.3daistudio.com/processing/rigging) - auto-rig characters before export Don't have a model yet? Generate one with [Image to 3D](https://www.3daistudio.com/ImageTo3D) or [Text to 3D](https://www.3daistudio.com/TextTo3D) and import it into Maya in minutes. ## Frequently Asked Questions ### How do I import an AI-generated 3D model into Maya? Go to File > Import in Maya, set the file type filter to FBX, and select the file you downloaded from 3D AI Studio. The model appears in your viewport with materials attached. FBX is the recommended format because Maya's Arnold shader reads FBX materials automatically. ### Do AI model textures work with Arnold rendering in Maya? Yes. FBX files from 3D AI Studio include PBR textures that connect to Arnold's aiStandardSurface shader automatically on import. Open the Hypershade to see the material connections for Base Color, Normal, Roughness, and Metalness. You can render immediately or tweak shader values to match your scene lighting. ### Can I rig an AI-generated character in Maya? You have two options. The faster route is to use the Rigging tool in 3D AI Studio, which gives you a pre-rigged FBX with bones already bound to the mesh. Import that directly into Maya and start animating. Alternatively, import the unrigged model and build your own joint skeleton using Maya's rigging tools for full control. ### Should I use FBX or OBJ when importing from 3D AI Studio into Maya? Use FBX whenever you need materials, hierarchy, or animation data to come through. OBJ only carries geometry, so you would need to set up materials manually in the Hypershade. 3D AI Studio's FBX export preserves everything Maya needs, making it the better choice for most production workflows. --- # ZBrush Integration Source: https://docs.3daistudio.com/integrations/zbrush ## Recommended Format **OBJ** is the best format for ZBrush. It imports cleanly and ZBrush handles OBJ geometry natively. ## Import Steps 1. Download your model as OBJ from 3D AI Studio 2. Open ZBrush 3. Go to **File > Import** 4. Select your .obj file 5. The model loads as a subtool in your scene ## After Import - The model comes in as a single subtool without textures - To load textures, go to **Texture Map > Import** and select your texture files - If the model feels too small or large, use Tool > Deformation > Size to adjust ## Common Workflows ### Use as a Sculpting Base AI-generated models make great starting points for detailed sculpting: 1. Import your OBJ 2. Apply DynaMesh (Tool > Geometry > DynaMesh) to get uniform topology 3. Set resolution based on detail level needed (128-512 for most models) 4. Sculpt in your details - skin pores, armor scratches, fabric folds, whatever you need ### Retopologize for Production After sculpting, clean up the topology for game engines or rendering: 1. Use ZRemesher (Tool > Geometry > ZRemesher) for automatic retopology 2. Project your sculpted details onto the new topology 3. Generate UVs with UV Master 4. Export as OBJ for your target application ### Export Back to Other Tools After sculpting in ZBrush, export your model for use in game engines or renderers: - OBJ for general compatibility - FBX if you need to keep polygroups as material assignments - Use Multi Map Exporter to bake normal and displacement maps ## Tips - If the poly count from 3D AI Studio is too high for comfortable sculpting, use the [Remesh tool](https://docs.3daistudio.com/processing/remesh) to reduce it before downloading - ZBrush works best with clean, watertight OBJ files - Use Sculptris Pro mode if you want ZBrush to handle topology automatically while you sculpt - Enable symmetry (Transform > Activate Symmetry) for character work - Save iterations with layers for non-destructive sculpting Don't have a model yet? Generate one with [Image to 3D](https://www.3daistudio.com/ImageTo3D) or [Text to 3D](https://www.3daistudio.com/TextTo3D) and bring it into ZBrush for sculpting. ## Frequently Asked Questions ### How do I import an AI-generated model into ZBrush? Download your model as OBJ from 3D AI Studio, then go to File > Import in ZBrush and select the .obj file. The model loads as a subtool that you can start sculpting on immediately. OBJ is the cleanest format for ZBrush since it handles the geometry natively. ### Can I sculpt on top of an AI-generated 3D model? Yes, and it is one of the best uses for AI-generated models. Import the OBJ from 3D AI Studio, apply DynaMesh to create uniform topology, then sculpt in details like skin pores, scratches, or fabric folds. The AI model gives you a solid base shape so you can focus on the creative detailing work. ### What should I do if the AI model has too many polygons for ZBrush? Use the Remesh tool in 3D AI Studio to reduce the polygon count before downloading. This gives you a lighter mesh that is easier to work with in ZBrush. Alternatively, you can apply ZRemesher inside ZBrush after import, but starting with a reasonable poly count saves time. ### How do I export a sculpted AI model from ZBrush back to a game engine? After sculpting, use ZRemesher to create clean topology, then generate UVs with UV Master. Export as OBJ or FBX from ZBrush and use Multi Map Exporter to bake normal and displacement maps. The original 3D AI Studio model gives you the base, ZBrush adds the detail, and the baked maps carry that detail into your game engine at a fraction of the polygon cost. --- # API Source: https://docs.3daistudio.com/api-reference 3D AI Studio has a REST API that lets you generate 3D models, images, textures, and process meshes from your own code. Use it to build automated pipelines, integrate 3D generation into your app, or batch-process assets. ## What You Can Do | Endpoint | What it does | | --- | --- | | **3D Generation (Prism)** | Text to 3D and Image to 3D with Prism models | | **3D Generation (Hunyuan)** | Text to 3D and Image to 3D with Hunyuan 3.0 and 3.5 | | **3D Generation (Trellis 2)** | Image to 3D with Trellis 2 | | **3D Generation (Hitem3D)** | Image to 3D and Multiview to 3D (front, back, left, right views) with Hitem3D 3.0 | | **Image Generation** | Generate and edit images with Gemini, Flux, and more | | **Texturing** | AI texture generation for 3D models | | **Mesh Tools** | Remesh, format conversion, topology optimization | | **3D Printing** | Prepare models for 3D printing | All endpoints return async results. You submit a job, get a task ID, and poll for completion. ## Getting Started 1. Go to the [API Dashboard](https://www.3daistudio.com/Platform/API) and continue with your 3D AI Studio account (two-factor authentication required — set it up in Settings → Passkeys & 2FA) 2. Generate an API key 3. Use the key in your requests as a Bearer token 4. Check the [full API documentation](https://www.3daistudio.com/Platform/API/Documentation) for endpoints, parameters, and code examples ## Quick Example ```bash # Generate a 3D model from text curl -X POST https://api.3daistudio.com/v1/generate \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{"prompt": "a medieval treasure chest", "model": "prism-3.1"}' ``` For full request/response schemas, authentication details, webhook setup, and code samples in Python, JavaScript, and cURL, see the complete documentation on the main site. ## Full Documentation The complete API reference with all endpoints, parameters, code examples, and interactive testing lives on the main 3D AI Studio platform: [Open API Documentation](https://www.3daistudio.com/Platform/API/Documentation) [Open API Dashboard](https://www.3daistudio.com/Platform/API) (manage keys, credits, billing) ## API Credits API calls use the same credit system as the web app. The cost per call depends on the endpoint and settings. Credit costs are documented per endpoint in the [API docs](https://www.3daistudio.com/Platform/API/Documentation). ## Frequently Asked Questions ### How do I generate a 3D model using the 3D AI Studio API? Send a POST request to the generation endpoint with your API key and either a text prompt or an image. The 3D AI Studio API returns a task ID, and you poll for the result until the model is ready for download. Full code examples in Python, JavaScript, and cURL are available in the API documentation. ### How much does it cost to use the 3D AI Studio API? The API uses the same credit system as the web app, so prices per generation match what you see on the dashboard. Credit costs vary by endpoint and AI model - for example, Prism 3.1 costs 35 credits per generation. You manage credits and billing through the API Dashboard at 3daistudio.com/Platform/API. ### Can I batch-generate multiple 3D models through the API? Yes. The 3D AI Studio API is async, so you can submit multiple generation jobs in parallel and poll each one for completion. This makes it practical to build automated pipelines, generate assets in bulk, or integrate 3D generation into your own application's workflow. ### What can I do with the 3D AI Studio API besides generating models? The API covers much more than just 3D generation. You can generate and edit images, apply AI textures to models, run remeshing and format conversion, and prepare models for 3D printing. Every tool available on the 3D AI Studio web platform has a corresponding API endpoint. --- # Troubleshooting Guide Source: https://docs.3daistudio.com/troubleshooting Most problems are fixed by refreshing the page, checking the [Dashboard](https://www.3daistudio.com/Dashboard) for the finished result, or switching to an up-to-date Chrome, Edge, Firefox or Safari. If a generation fails, the credits are refunded to your account automatically. | Problem | Quick fix | | --- | --- | | Generation failed | Refresh and try again. Credits are refunded automatically | | Generation seems stuck | Check the Dashboard, results finish in the background | | Model looks low quality | Use a cleaner input image or a more specific prompt | | Upload fails | Check the file format and the size limit shown in the tool | | 3D viewer is black | Enable hardware acceleration and update graphics drivers | | Textures missing after download | Use GLB, or keep OBJ, MTL and textures in one folder | Using 3D AI Studio from Claude, ChatGPT, Cursor or another AI assistant? See [MCP Troubleshooting](https://docs.3daistudio.com/mcp/troubleshooting). ## Do I get my credits back when a generation fails? Yes. When a generation or processing step fails, the credits are refunded to your account automatically. If the balance does not update right away, refresh the page or check again after a few minutes. ## Why did my generation fail? The usual causes are a dropped internet connection, an input the model cannot use, or temporary high server load. Refresh the page and run the generation again; most failures succeed on the second try. 1. **Check your connection.** Refresh the page and close bandwidth-heavy apps. 2. **Check the input.** Use a supported image (PNG, JPG or WebP) with one clear subject, or a short, focused text prompt. 3. **Wait and retry.** During high load, try again after 5-10 minutes. 4. **Contact support** at [support@3daistudio.com](mailto:support@3daistudio.com) if it keeps failing. ## Why is my generation taking so long? Generation time depends on the model: from about one minute (Tripo P1) to 3-5 minutes for the default Prism 3.1 and up to 10-12 minutes for the slowest models. Higher texture and mesh quality add time. Generations keep running when you close the page and appear in your [Dashboard](https://www.3daistudio.com/Dashboard) when done. If it is far beyond the usual time for the model: 1. Refresh the page 2. Check the Dashboard for the finished model 3. Clear the browser cache, or try another browser or device Times per model are listed in [Image to 3D](https://docs.3daistudio.com/3d-generation/image-to-3d) and [Text to 3D](https://docs.3daistudio.com/3d-generation/text-to-3d). ## Why does my 3D model look low quality? The input matters most. A sharp, well-lit image of one object on a plain background, or a specific prompt naming materials and style, gives clearly better models than blurry photos or vague prompts. - **Image input:** use at least 1024x1024 pixels, even lighting and a simple background. Remove busy backgrounds first in [Image Studio](https://docs.3daistudio.com/image-studio/utilities). - **Text input:** describe one object with its materials and style, for example "weathered wood", "metallic finish". See [Prompt Tips](https://docs.3daistudio.com/3d-generation/prompt-tips). - **Settings:** enable PBR materials and choose a higher texture quality. - **Hidden sides wrong:** use a multi-view model with photos from several angles. - **Messy topology:** clean it up with the [Remesh tool](https://docs.3daistudio.com/processing/remesh). - **Best results overall:** follow the [recommended workflow](https://docs.3daistudio.com/3d-generation/recommended-workflow): perfect the image first, then convert it. ## Why won't my file upload? The file is usually in an unsupported format, larger than the tool's limit, or damaged. Check the format first, then try a smaller or freshly exported file. | Upload | Supported formats | | --- | --- | | Images (Image to 3D, Image Studio) | PNG, JPG, JPEG, WebP | | 3D models | GLB, GLTF, FBX, OBJ, STL | | Texture maps | PNG, JPG | - **Size limits** depend on the tool and are shown in the upload area. For example, the [Texture Generator](https://docs.3daistudio.com/textures/texture-generator) accepts models up to 120MB (200MB with Hunyuan). - **Damaged files:** export the file again from the source software and test with a smaller file. - **Converting 3D files:** export from Blender, Maya or 3ds Max as FBX or OBJ with textures included, or use the free in-browser [3D file converter](https://www.3daistudio.com/convert). ## Why is the 3D viewer black or not loading? The 3D viewer needs WebGL. A black or empty viewer almost always means WebGL or hardware acceleration is disabled, or the graphics driver is outdated. 1. **Enable hardware acceleration** in your browser settings and restart the browser 2. **Update your graphics drivers** and restart your computer 3. **Close graphics-heavy apps** such as games or video editors 4. **Try Chrome or Edge**, which give the best viewer performance 5. **Try another device** if it still fails ## Which browsers are supported? Use the latest version of Chrome, Edge, Firefox or Safari with JavaScript and WebGL enabled. Internet Explorer is not supported. | Browser | Support | | --- | --- | | Chrome | Recommended, best performance | | Edge | Recommended on Windows | | Firefox | Supported | | Safari | Supported on Mac | | Internet Explorer | Not supported | Allow pop-ups and cookies for 3daistudio.com, and disable ad blockers on the site if downloads or login misbehave. ## Why won't my model download? Downloads are usually blocked by browser permissions, a download manager or a dropped connection. Allow downloads for 3daistudio.com, check free disk space and try again on a stable connection. - Check the browser's download permissions and pop-up blocker - Disable download manager extensions that intercept files - Try a different format, for example GLB instead of FBX - To download many files at once, use [ZIP Export](https://docs.3daistudio.com/utilities/zip-export) ## Why are textures missing after download, or why does the model look different? Missing textures usually come from OBJ files whose .mtl and image files were separated. Download GLB, which keeps everything in one file, or keep all OBJ files in one folder. A model that looks different from the web preview is normal: your 3D software uses its own lighting. Switch to a material or PBR preview mode there. - Use **GLB** for a single file with textures embedded - For OBJ, keep the .obj, .mtl and textures together, or download **OBJ ZIP** - Use **FBX** for game engines and **STL** for 3D printing - Import guides: [Blender](https://docs.3daistudio.com/integrations/blender), [Unity](https://docs.3daistudio.com/integrations/unity), [Unreal Engine](https://docs.3daistudio.com/integrations/unreal), [Godot](https://docs.3daistudio.com/integrations/godot) ## What do the error messages mean? | Message | Meaning | Fix | | --- | --- | --- | | **Authentication Failed** | Your session expired | Log out and back in, clear cookies | | **Quota Exceeded** | Not enough credits | Check your balance, top up or upgrade | | **File Too Large** | The file is over the tool's limit | Compress or reduce the file and upload again | | **Invalid File Format** | Unsupported or damaged file | Convert to a supported format and re-export | | **Server Error 500** | Temporary server problem | Retry in 5-10 minutes, contact support if it persists | ## Why is the site slow on my computer? 3D previews use your graphics card and memory. Close unused tabs and heavy apps, disable unnecessary extensions, and use a device with at least 8GB of RAM and a dedicated graphics card for the smoothest experience. | | RAM | Graphics | | --- | --- | --- | | Minimum | 4GB | Modern graphics card with WebGL | | Recommended | 8GB or more | Dedicated graphics card | | Best | 16GB or more | Modern GPU | Very heavy models can make the viewer choppy. Reduce them with the [Remesh tool](https://docs.3daistudio.com/processing/remesh). ## How do I contact support? Email [support@3daistudio.com](mailto:support@3daistudio.com) or ask in the [Discord community](https://discord.gg/uA3u6J7EDC). Include the following so the team can help quickly: 1. **Browser and device:** browser and version, operating system, RAM and graphics card 2. **The problem:** exact error message, what you did before it happened, screenshots if possible 3. **Account details:** your account email, the approximate time, and the tool you used Step-by-step videos are in [Video Tutorials](https://docs.3daistudio.com/learning). --- # FAQ Source: https://docs.3daistudio.com/faq ## Pricing & credits ### Is 3D AI Studio free to use? You can sign up and get free credits to try everything out. No credit card required. Free credits are enough for several 3D model generations. After that, paid plans start at $19/month for 1,000 credits. ### How do credits work? Every generation costs a set number of credits. Image to 3D and Text to 3D cost 20-60 credits depending on the AI model you pick. Image generation costs 2-8 credits. Some tools like SVG to 3D, 3D Text, and the 3D Viewer are completely free. Credit costs are always shown before you generate. ### What are the pricing plans? Basic is $19/month for 1,000 credits. Studio is $29/month for 3,500 credits. Pro is $149/month for 18,000 credits. There's also a one-time pack at $34 for 2,000 credits that never expire. All plans include access to every tool. [View pricing](https://www.3daistudio.com/Pricing) ### Do unused credits roll over? Monthly credits do not roll over between billing cycles. One-time credit packs never expire. ### Can I cancel my subscription? Yes, anytime from your account settings. Your access continues until the end of the billing period. No cancellation fees. ## Creating 3D models ### Do I need 3D modeling experience? No. Everything runs in the browser. You type a description or upload an image, and the AI generates a 3D model. No Blender, Maya, or any other software needed. You can preview and export directly. ### Which AI model should I use for 3D generation? Prism 3.1 is the best all-around choice (35 credits, 3-5 minutes). For game-ready assets, Tripo P1 (60 credits, 60-90 seconds) or Tripo P2 (70 credits, quad meshes). For highest quality, Hunyuan 3.1 Pro (40 credits). Prism 3.1 also has a Latest (experimental) texture model setting - results can vary, so stick with the Standard texture model for production work. [Full model comparison](https://docs.3daistudio.com/3d-generation/image-to-3d) ### What image formats can I upload for Image to 3D? PNG, JPG, JPEG, and WebP. For best results, use a clear image with one subject, simple background, and good lighting. At least 1024x1024 pixels recommended. ### How do I get better results? Start in Image Studio to generate or edit a clean reference image first, then convert to 3D. This two-step workflow gives much better results than Text to 3D alone. [Recommended Workflow guide](https://docs.3daistudio.com/3d-generation/recommended-workflow) ### Can I rig and animate my models? Yes. Use the [Rigging tool](https://docs.3daistudio.com/processing/rigging) (15-20 credits) to add a skeleton to character models. Supports biped, quadruped, and other creature types. After rigging, apply animations from 12 presets (walk, run, dance, etc.) for 15 credits each. ### What browsers are supported? Any modern browser with WebGL support. Chrome, Firefox, Safari, and Edge all work. No plugins or downloads needed. Works on desktop and mobile. ### Can I use 3D AI Studio from Claude, ChatGPT or Cursor? Yes. With the MCP connector (included in every paid plan) you connect Claude, ChatGPT, Cursor, Claude Code, Codex and other AI assistants to your account and generate 3D models and images right from the chat. Generations use your normal credits and land in your dashboard. [MCP Quick Start](https://docs.3daistudio.com/mcp/quick-start) Have a question not covered here? Contact [support@3daistudio.com](mailto:support@3daistudio.com) or join the [Discord](https://discord.gg/uA3u6J7EDC). ## Files, storage & privacy ### What file formats can I export? GLB (recommended default), FBX (game engines), OBJ (universal), STL (3D printing), USDZ (Apple AR), BLEND (Blender), 3MF (color 3D printing), Collada, and PLY. You can convert between formats using the Change Format action (1 credit) or the [Remesh tool](https://docs.3daistudio.com/processing/remesh). [Full format guide](https://docs.3daistudio.com/export-formats) ### Can I 3D print the models? Yes. Use [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing): it makes the model watertight, sets the size in millimeters and downloads an STL for single-color printing or a 3MF for multi-color printers (Bambu Lab AMS, Prusa MMU). The [Remesh tool](https://docs.3daistudio.com/processing/remesh) can reduce the polygon count of heavy models first. ### How long are my 3D models stored? Generated 3D models stay in your library for 6 months on the free plan and 12 months on paid plans, counted from when they were created. Any model can be extended to 3 years for 1 credit from its details panel, and every file shows the exact date it's kept until. Your inputs, settings and previews are kept regardless. [How model storage works](https://docs.3daistudio.com/utilities/model-storage) ### Is my data secure? Yes. Your models and images are stored securely and are not shared or used for training. You can delete your data at any time from your account settings. --- # What Is Rigging in 3D Animation? Source: https://docs.3daistudio.com/concepts/what-is-rigging Rigging is the step that turns a static 3D model into something you can animate. A rigger places a skeleton of connected bones inside the model and binds the mesh to it, so moving a bone bends the surface around it, the same way an arm moves when you bend the elbow. | | | | --- | --- | | **What it adds** | A skeleton (bones and joints) plus skin weights | | **Why** | To pose and animate characters, creatures and machines | | **Comes after** | Modeling and texturing | | **Comes before** | Animation | | **Typical tools** | Blender, Maya, Mixamo, AI auto-rigging | ## What does a rig consist of? A rig has three layers: a **skeleton** that defines how the model can move, **skin weights** that decide which part of the mesh follows which bone, and optional **controls** that make the skeleton easy for an animator to pose. - **Skeleton (armature):** a hierarchy of bones, usually starting at the hips. Moving a parent bone, like the upper arm, moves its children, like the forearm and hand. - **Skin weights:** for every vertex of the mesh, how strongly each nearby bone pulls on it. Smooth weights around joints are what keep elbows and knees from folding like paper. - **Controls:** handles and constraints on top of the skeleton, for example a foot control that keeps the foot planted while the body moves. - **Facial rigs:** faces often use extra bones or blend shapes (shape keys) for expressions and lip sync. ## How does rigging work in animation? Animators never move the mesh directly. They move the rig, and the skin weights carry that motion to the surface. An animation is a sequence of rig poses over time, which is why one rig can play many animations and why animations can be moved between characters with the same skeleton. Two ways of posing a limb are common: | | Forward kinematics (FK) | Inverse kinematics (IK) | | --- | --- | --- | | How you pose | Rotate each bone in turn, from shoulder to hand | Move the hand or foot, and the joints in between follow | | Good for | Swinging arms, flowing motion | Feet on the ground, hands holding objects | ## What types of rigs are there? Rigs follow the body plan of the model. A humanoid rig suits people and human-like characters, while animals, birds and snakes need skeletons shaped for how they move. - **Biped (humanoid):** two legs and two arms. Game engines have dedicated humanoid systems, such as Unity's Humanoid rig type - **Quadruped:** four-legged animals like dogs, horses or dragons - **Avian:** birds and other winged creatures - **Serpentine:** snakes, tails and tentacles, built from long chains of bones - **Mechanical:** robots and vehicles, where parts are usually attached to bones rigidly instead of being skinned ## How do you rig a 3D model? You can rig by hand in software like Blender or Maya, or let an automatic rigger do it. Manual rigging gives full control but takes time and experience. Auto-rigging places the bones and paints the weights for you in minutes. 1. **Prepare the model:** one connected mesh in a neutral pose, ideally a T-pose or A-pose with the arms away from the body 2. **Build the skeleton:** place joints at the natural pivot points, like shoulders, elbows, hips and knees 3. **Bind the mesh:** calculate skin weights so each vertex follows the right bones 4. **Fix the weights:** check extreme poses and smooth any areas that stretch or collapse 5. **Add controls:** optional IK handles and constraints for animators The [AI rigging tool](https://www.3daistudio.com/Tools/RiggingTool) in 3D AI Studio does steps 2 to 4 automatically. It rigs bipeds, quadrupeds, birds and serpents in about 1-2 minutes for 15-20 credits, offers 12 preset animations, and exports GLB or FBX for Unity and Unreal, with Mixamo-compatible bones on Prism rigs. The [rigging guide](https://docs.3daistudio.com/processing/rigging) walks through every setting. ## Do I need to rig a model for 3D printing? No. Rigging only matters when a model has to move. For a print, a static model in a good pose is enough. A rig can still help you pose a character before you export it, but the printed file itself contains no bones. ## Frequently Asked Questions ### What is the difference between rigging and skinning? Rigging is the whole process of building the skeleton and its controls. Skinning, also called binding or weight painting, is the part of rigging that connects the mesh to the bones by assigning skin weights. A rig without skinning has bones that move, but the model stays still. ### Can AI rig a 3D model automatically? Yes. AI auto-riggers detect the body shape, place the joints and paint the skin weights without manual work. Results are best on clean, single-mesh characters in a neutral pose. For unusual creatures or production rigs with complex controls, riggers still refine the result by hand. --- # What Is a Normal Map? Source: https://docs.3daistudio.com/concepts/what-is-a-normal-map A normal map is a texture that makes a 3D surface look more detailed than it really is. Each pixel stores a direction instead of a color, and the renderer uses that direction to light the surface, so bumps, grooves and seams appear without adding any polygons. | | | | --- | --- | | **Stores** | The direction the surface faces at each pixel, as RGB | | **Changes** | How light falls on the surface, not its shape | | **Typical look** | Light blue and purple | | **Part of** | PBR materials, next to base color, roughness and metalness | | **Used in** | Games, real-time viewers, renders | ## How does a normal map work? Lighting depends on which way a surface faces, its "normal". A normal map replaces the flat face's normal with a different direction at every pixel. Light then hits each pixel as if it sat on a tiny slope, which the eye reads as real detail like pores, scratches or panel lines. The direction is encoded in the color channels: red for left and right (X), green for up and down (Y), and blue for straight out of the surface (Z). The geometry stays exactly as it was, which is why normal maps are so cheap to render. ## Why are normal maps blue and purple? Most normal maps are tangent-space maps, where a pixel pointing straight out of the surface is stored as RGB (128, 128, 255). That value is a light blue-purple, and since most of a surface points roughly straight out, most of the map has that color. Slopes show up as pinker or greener tints. Object-space normal maps store directions relative to the whole model instead and look like a rainbow of colors. They are less common because they don't work on deforming characters. ## What is the difference between OpenGL and DirectX normal maps? The two conventions flip the green channel. A map made for one looks inverted in the other, with bumps that appear to point inward. Flip the green channel when moving a map between them. | Convention | Green channel | Used by | | --- | --- | --- | | **OpenGL** (Y+) | Up is green | Blender, Unity, Godot | | **DirectX** (Y-) | Down is green | Unreal Engine, 3ds Max | ## Normal map vs bump map vs displacement map All three add surface detail, but only displacement changes the actual shape. Normal maps are the standard for real-time 3D because they look detailed and cost almost nothing. | | Normal map | Bump map | Displacement map | | --- | --- | --- | --- | | Stores | Direction (RGB) | Height (grayscale) | Height (grayscale) | | Changes geometry | No | No | Yes | | Silhouette | Unchanged | Unchanged | Changes | | Cost | Very low | Very low | High, needs dense geometry | ## Where do normal maps come from? Normal maps are usually baked from a detailed high-poly model onto a simpler low-poly version, generated from a photo or texture, or produced together with the other maps of an AI-textured model. - **Baking:** the detail of a high-poly sculpt is projected onto a [low poly](https://docs.3daistudio.com/concepts/what-is-low-poly) model, which is how game assets keep their detail - **From an image:** a [normal map generator](https://www.3daistudio.com/Tools/PBRMapGenerator) estimates surface direction from a photo or texture - **With AI texturing:** the [AI texture generator](https://www.3daistudio.com/TextureGenerator) creates base color, roughness, metalness and normal maps for a whole model at once ## How do I use a normal map? Plug it into the normal input of your material and make sure the engine treats it as non-color data. Then check the green channel convention. - **Blender:** Image Texture node with color space **Non-Color**, into a **Normal Map** node, into the Normal input of the Principled BSDF - **Unity:** set the texture's type to **Normal map** in the import settings, then assign it to the material's Normal Map slot - **Unreal Engine:** import with the **Normalmap** compression setting, and flip the green channel if the map was made in OpenGL style - **Godot:** enable **Normal Map** in the StandardMaterial3D and assign the texture ## Frequently Asked Questions ### Why does my normal map look wrong or inverted? The two most common causes are a flipped green channel (an OpenGL map in a DirectX engine or the other way round) and the texture being read as color instead of non-color data. Fix the color space first, then flip the green channel if bumps still point inward. ### Do I need a normal map for 3D printing? No. A printer only uses the geometry, so normal map detail disappears in a print. To keep painted or fine detail physically, it has to become real geometry, for example with DetailForge in [Export for 3D Printing](https://docs.3daistudio.com/utilities/export-for-3d-printing). --- # What Does Low Poly Mean? Source: https://docs.3daistudio.com/concepts/what-is-low-poly Low poly means a 3D model made of relatively few polygons, the flat faces that form every 3D surface. The term describes two things: models kept light so they run fast in games, AR and the web, and an art style that shows its flat, faceted shapes on purpose. | | | | --- | --- | | **Polygon** | A flat face, usually a triangle or a quad | | **Low poly** | Few polygons for the job, often hundreds to a few thousand | | **High poly** | Many polygons, often hundreds of thousands or millions | | **Why** | Performance, small files, or a stylized look | | **Made by** | Modeling by hand, retopology, decimation or AI | ## How many polygons is low poly? There is no fixed number. Low poly is relative to the platform and the object: a character for a mobile game might use a few thousand triangles, while the same character in a PC game could use tens of thousands and still count as optimized. | Use | Typical face count | | --- | --- | | Mobile games, web viewers | 1,000-5,000 | | PC and console games | 5,000-30,000 | | Film, renders, 3D printing | 10,000-200,000 | Game engines usually count triangles, because every quad is split into two triangles for rendering. AI-generated and scanned models often start at hundreds of thousands of triangles, which is why they are reduced before use in real-time projects. ## What is the difference between low poly and high poly? A high poly model carries every detail in its geometry. A low poly model keeps only the shape and gets its fine detail from textures, especially [normal maps](https://docs.3daistudio.com/concepts/what-is-a-normal-map) baked from the high poly version. | | Low poly | High poly | | --- | --- | --- | | Polygon count | Low | High | | Detail comes from | Textures and normal maps | Geometry | | Performance | Fast | Slow in real time | | File size | Small | Large | | Typical use | Games, AR/VR, web | Sculpting, film, renders, baking source | ## What is the low poly art style? As an art style, low poly embraces visible facets: flat-shaded faces, simple shapes and often solid colors instead of detailed textures. It is popular in indie games, illustrations and motion graphics because it reads clearly, looks clean and renders quickly. ## How do you make a low poly model? Model it by hand with few polygons from the start, or reduce an existing detailed model. Reducing works in two ways: decimation removes polygons automatically, and retopology rebuilds a clean new mesh over the original shape. - **Modeling by hand:** start from simple shapes in Blender or Maya and add only the edges you need - **Decimation:** a quick automatic reduction, such as Blender's Decimate modifier. Fast, but the edge flow is messy - **Retopology:** a new, clean mesh that follows the original surface, with textures baked across. Better for animation and texturing [Smart Low Poly](https://docs.3daistudio.com/processing/smart-low-poly) in 3D AI Studio does AI retopology on any model: it rebuilds the mesh as 500-20,000 triangles or 500-10,000 quads and bakes the original textures onto it. To create a low poly model from scratch, the [AI low poly generator](https://www.3daistudio.com/UseCases/LowPoly) produces low poly models and characters directly. ## Frequently Asked Questions ### Is low poly the same as low quality? No. A good low poly model uses its polygons where they matter and gets fine detail from textures and normal maps, so it can look excellent while running fast. Low poly is about efficiency, or a deliberate style, not about a worse result. ### What is polycount? Polycount is the number of polygons in a model, usually counted in triangles for games. It is the main number to watch when a model has to run in real time, together with texture size and the number of materials. --- # Video Tutorials Source: https://docs.3daistudio.com/learning Watch tutorials covering the full platform. New videos are added regularly. ## Featured: Full Platform Walkthrough [Video: 3D AI Studio Full Guide by PixelArtistry](https://www.youtube.com/watch?v=MUp-dMf8mLU) *Complete walkthrough by PixelArtistry covering the full 3D AI Studio platform.* ## Quick Guides [Video: Image to 3D Tutorial](https://www.youtube.com/watch?v=OVKl7idiyIU) *Image to 3D - from photo to 3D model in minutes.* [Video: Text to 3D Tutorial](https://www.youtube.com/watch?v=Y05LgsWPjoY) *Text to 3D - describe what you want and generate a 3D model.* [Video: AI for Game Development](https://www.youtube.com/watch?v=L5emNlNYVfY) *AI for game development - full pipeline from generation to game engine.* [Video: AI Models for 3D Printing](https://www.youtube.com/watch?v=ykMGnz5CR2I) *AI models for 3D printing - which ones work and how to prepare them.* ## More Tutorials For a full library of step-by-step video tutorials covering every tool, visit the [Learning Studio](https://www.3daistudio.com/Tools/Tutorials) on the main platform. [Open the Video Tutorial Library](https://www.3daistudio.com/Tools/Tutorials) ## Frequently Asked Questions ### Are there video tutorials for 3D AI Studio? Yes. 3D AI Studio has a full library of video tutorials covering everything from getting started to advanced workflows like game dev pipelines and 3D printing preparation. The featured walkthrough on this page covers the entire platform in one video, and shorter guides focus on specific tools like Image to 3D and Text to 3D. ### Where can I learn AI 3D modeling as a complete beginner? Start with the Getting Started guide and the full platform walkthrough video on this page. 3D AI Studio is designed for beginners, so you do not need prior 3D modeling experience. The video tutorials walk you through each step visually, from signing up to downloading your finished model. ### Does 3D AI Studio have tutorials for game development workflows? Yes. The "AI for Game Development" tutorial on this page covers the full pipeline from generating a model in 3D AI Studio to importing it into a game engine. It includes tips on format selection, polygon optimization, and material setup for real-time rendering. ### Where can I find more tutorials beyond this page? Visit the Learning Studio at 3daistudio.com/Tools/Tutorials for the complete video library with step-by-step guides for every tool. New tutorials are added regularly as 3D AI Studio releases new features. You can also find community-made tutorials and tips in the Discord server. --- # Community Source: https://docs.3daistudio.com/community Connect with the 3D AI Studio community across our social media platforms. Join thousands of creators, share your work, get support, and stay updated with the latest features and announcements. ## Discord Community Join our active Discord server where creators collaborate, share projects, and help each other master 3D AI generation. Get real-time support, participate in community challenges, and connect with fellow 3D enthusiasts. **[Join the 3D AI Studio Discord Server](https://discord.gg/uA3u6J7EDC)** 🎮 Our Discord community is the perfect place to: - Get instant help and feedback on your models - Share your latest creations and get inspired by others - Stay updated with new features and platform updates - Participate in community events and challenges - Connect with creators at all skill levels ## Follow Us on Social Media Stay connected and get the latest updates across all our social platforms: **Twitter** 🐦 Follow us [@3daistudio](https://twitter.com/3daistudio) for quick updates, feature announcements, and community highlights. Get the latest news and join conversations about 3D AI generation. **Instagram** 📸 Follow us [@3daistudio](https://instagram.com/3daistudio) to discover amazing community creations, behind-the-scenes content, and visual inspiration. See what's possible with 3D AI Studio through stunning model showcases. ## Community Guidelines We maintain a welcoming and supportive environment for all creators. Whether you're just starting with 3D AI generation or you're a seasoned professional, our community is here to help you succeed and grow your skills. Join us today and become part of the growing 3D AI Studio community! New to 3D AI Studio? [Get started for free](https://www.3daistudio.com) -- try [Image to 3D](https://www.3daistudio.com/ImageTo3D), [Text to 3D](https://www.3daistudio.com/TextTo3D), or explore the [Image Studio](https://www.3daistudio.com/ImageGeneration) with free credits on signup. ## Frequently Asked Questions ### Does 3D AI Studio have a Discord community? Yes. The 3D AI Studio Discord server is the main community hub where creators share models, get feedback, ask for help, and stay updated on new features. Join at discord.gg/uA3u6J7EDC to connect with thousands of other 3D AI enthusiasts. ### Can I share my AI-generated 3D models with the community? Absolutely. The 3D AI Studio Discord has channels for sharing your creations and getting feedback from other users. You can also post your work on social media and tag @3daistudio on Twitter or Instagram. Community challenges and events give you a chance to showcase your best work. ### Where do I get help if I am stuck on something in 3D AI Studio? The Discord community is the fastest way to get help from both other users and the support team. Post your question with screenshots or details about what you are trying to do, and you will usually get a response quickly. For account or billing issues, email support@3daistudio.com directly. ### How do I stay updated on new 3D AI Studio features? Follow @3daistudio on Twitter for quick announcements and feature updates. The Discord server also has a dedicated announcements channel where the 3D AI Studio team posts about new AI models, tools, and platform improvements as soon as they launch.