---
title: "VS Code MCP Setup (GitHub Copilot)"
description: "Add the 3D AI Studio MCP server to VS Code with mcp.json or the MCP: Add Server command, sign in, and generate 3D models and images in Copilot agent mode."
source: https://docs.3daistudio.com/mcp/setup/vs-code
---

# VS Code MCP Setup (GitHub Copilot)

Add the 3D AI Studio MCP server to VS Code with mcp.json or the MCP: Add Server command, sign in, and generate 3D models and images in Copilot agent mode.

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 **Steps look different?** VS Code and Copilot change their MCP support frequently, so the steps below can go out of date.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.

---

> Documentation index for AI agents: https://docs.3daistudio.com/llms.txt