Fusion-360

This MCP server enables Autodesk Fusion 360 to be controlled by AI agents through the Aura Friday MCP-Link infrastructure. The server exposes Fusion integration tools that allow remote agents to execute generic Fusion API calls, run Python code inside Fusion with full access to the Fusion API (adsk.core, adsk.fusion, etc.), and leverage the built-in MCP toolset (Browser, SQLite, OpenRouter, Desktop automation, and more) to perform complex design-and-automation workflows. Agents can send API paths and code blocks to manipulate designs, query and store data in SQLite, automate browser tasks, and present results to users via Fusion popups. The included examples demonstrate how to execute API paths, run arbitrary Python, and utilize the MCP tool suite to automate tasks inside Fusion. To start, ensure the MCP server is running and Fusion is configured to connect to the MCP-Link server; the add-in will auto-connect on startup and log connection status in the Fusion TEXT COMMANDS window.

Prerequisites:

• Autodesk Fusion 360 installed
• Access to the Aura Friday MCP-Link server (downloadable from the official releases)
• Python environment available (if using the uvx workflow)
• Optional: Node.js if you plan to use any Node-based tooling in your workflow

Installation steps:

1. Prepare Fusion add-in installation:
   • Install the Autodesk Fusion add-in that exposes the MCP integration (as provided by Aura Friday).
   • Clone or download the Fusion-360-MCP-Server repository if you have the source.
2. Install the MCP-Link server (server that coordinates MCP tools):
   • Download and run the MCP-Link server from Aura Friday releases: follow the official instructions for your OS.
3. Install and start the Fusion MCP server package:
   • If using uvx (Python tooling), install the package via:

     uvx aurafriday-fusion-360-mcp-server
     

   • If using an alternative method, follow the package’s installation instructions and ensure the server exposes the expected endpoints.
4. Configure and connect:
   • Ensure Fusion is configured to auto-connect to the MCP server on startup (as described in the readme).
   • Verify connectivity by checking the TEXT COMMANDS window in Fusion for a successful MCP connection message.
5. Validate operation:
   • Open Fusion, run a simple example (e.g., a basic API call via fusion360.execute) to confirm the server can receive and execute commands.

Prerequisites recap:

• Fusion 360 installed and able to load the provided add-in
• Access to the Aura Friday MCP-Link server
• A working Python/uvx environment if you’re using the Python-based installation path
• Network access between Fusion and the MCP-Link server

Tips and notes:

• Environment variables: Set MCP_SERVER_URL to point to your MCP-Link server; MCP_AUTH_TOKEN may be required for secure setups.
• Logs: Check Fusion's TEXT COMMANDS window for connection status and error messages when the add-in attempts to connect to the MCP server.
• Python integration: The server supports executing arbitrary Python within Fusion, including access to adsk.core, adsk.fusion, and SQLite. Use caution with code execution—include proper error handling and sandboxing where possible.
• Tools availability: The MCP toolset includes Browser automation, SQLite, Python, OpenRouter, HuggingFace, Context7, Desktop automation, User popups, Remote connectors, and more. These work out-of-the-box with the MCP server and do not require additional configuration beyond connectivity.
• Documentation access: The server can fetch AI models, docs, and best practices via integrated tools. Use the get_api_documentation, get_online_documentation, and get_best_practices operations to discover available capabilities.
• Updates: If you enable automatic updates, ensure code-signing and signature verification are in place to avoid tampering.
• Compatibility: This server is designed to enable Fusion as a remote-controlled tool; ensure your agent workflows respect Fusion’s licensing and usage policies.