Images & Uploads
How to give your AI assistant a photo, sketch or reference image for 3D generation or editing: the inline upload card, the upload page, links and library images.
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.
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.
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 for more on the image first approach.
Frequently Asked Questions
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.
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.
Only you. Uploads are tied to your account and are only used for the generations you ask for. See Security & Privacy.
What You Can Do
Every tool in the 3D AI Studio MCP connector: 3D generation models, image generation and editing, low poly, UVs, format conversion and PBR materials, with costs.
Credits & Limits
How the 3D AI Studio MCP connector spends credits: website prices, personal or team credits, confirmations for large requests, refunds, rate limits and plan needs.