What is Fusion Mcp Server

Fusion-MCP-Server is a remote control solution for Autodesk Fusion 360 that utilizes a client-server architecture. It allows external applications to interact with Fusion 360, enabling command monitoring, remote execution, and model data retrieval for automation workflows.

Use cases

Use cases for Fusion-MCP-Server include automating repetitive design tasks, integrating Fusion 360 with other applications for enhanced workflows, and enabling remote collaboration on design projects.

How to use

To use Fusion-MCP-Server, first install the server and client by cloning the repository and ensuring Python 3.6+ is installed. Start the server by running ‘python server.py’ in the command prompt. Then, connect Fusion 360 to the server by adding the Fusion 360 add-in and running it.

Key features

Key features of Fusion-MCP-Server include JSON socket communication, a modular design with a Python server and client library, and integration with Fusion 360 via an add-in. It supports remote execution of commands and retrieval of model information.

Where to use

Fusion-MCP-Server can be used in fields such as engineering, product design, and automation where Fusion 360 is utilized for CAD modeling and design automation.

Fusion 360 MCP Server

Master Control Program (MCP) server for Autodesk Fusion 360 that enables remote control and automation of Fusion 360 operations.

Overview

This project implements a client-server architecture that allows external applications to interact with Fusion 360. The server acts as a bridge between clients and the Fusion 360 API, enabling remote execution of commands and retrieval of model information.

Components

The project consists of three main components:

1. MCP Server (server.py): A standalone Python server that listens for connections from clients and communicates with Fusion 360.

2. MCP Client (client.py): A Python client library that connects to the MCP server and provides methods for sending commands and receiving responses.

3. Fusion 360 Add-in (fusion360_mcp_addin.py): A Fusion 360 add-in that connects to the MCP server and provides the actual integration with the Fusion 360 API.

Installation

Server and Client

1. Clone this repository or copy the files to your desired location.

2. Make sure you have Python 3.6+ installed.

3. No additional Python packages are required for basic functionality as the implementation uses only standard library modules.

4. To enable LLM features, install the optional openai package:

   pip install openai
   

Fusion 360 Add-in

1. Copy the entire folder containing fusion360_mcp_addin.py, client.py, and the resources directory to your Fusion 360 AddIns directory. Ensure the resources/MCPIcon folder contains the required icon files (32x32-normal.png and 16x16-normal.png).
2. In Fusion 360, open the “Scripts and Add-ins” dialog (press Shift+S or find it in the “Design” workspace under “Utilities”).
3. On the “Add-ins” tab choose Load from my computer (“Load from Device” on some versions) and select this folder. Selecting the folder starts the add-in.
4. Click “Run” or enable “Run on Startup” to have it automatically load when Fusion 360 starts.

Usage

Starting the Server

1. Open a command prompt or terminal.
2. Navigate to the directory containing the server files.
3. Run the server:

python server.py

By default, the server will listen on 127.0.0.1:8080. You can modify the host
and port in the code if needed. Set the OPENAI_API_KEY environment variable if
you plan to use the LLM integration.

Connecting Fusion 360 to the Server

1. Start Fusion 360 and make sure the MCP add-in is running.
2. In Fusion 360, look for the “MCP Controls” panel.
3. Click the “Connect to MCP Server” button.
4. Enter the server host and port, then click OK.
5. If successful, you will see a confirmation message.

Using the Client

You can use the provided client implementation to connect to the server and interact with Fusion 360:

from client import MCPClient

# Create and connect the client
client = MCPClient('127.0.0.1', 8080)
if client.connect():
    # Get model information
    client.get_model_info()
    
    # Execute a Fusion 360 command
    client.execute_fusion_command('create_circle', {
        'center': [0, 0, 0],
        'radius': 10
    })
    
    # Disconnect when done
    client.disconnect()

Protocol

The server and clients communicate using a simple JSON-based protocol over TCP sockets. Each message is a JSON object with at least a type field indicating the message type.

Message Types

• fusion_command: Execute a command in Fusion 360
• get_model_info: Request information about the current model
• command_result: Response containing the result of a command execution
• model_info: Response containing model information
• llm_request: Request text generation from the configured LLM
• llm_result: Response containing LLM output

Extension

You can extend the server and add-in to support additional functionality:

1. Add new message types to the protocol
2. Implement handlers for these message types in the server
3. Add corresponding methods to the client
4. Implement the actual functionality in the Fusion 360 add-in

License

This project is licensed under the MIT License. See the LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Notes

• Fusion 360 (2601.1.34)MCP Add-in refuses to install with errors indicating more than one file.
  • Create a new add-in (python), edit, copy the fusion360_mcp_addin.py source to the new file, save, open the new file location and copy the client.py and resource folder to that location, restart Fusion 360.