Rhino

🚀 RhinoMCP - Rhino Model Context Protocol Integration

RhinoMCP connects Rhino with AI agents through the Model Context Protocol (MCP), enabling AI agents to directly control and interact with Rhino. This integration facilitates prompt-assisted 3D model design in Rhino 3D.

🚀 Quick Start

RhinoMCP integrates Rhino with AI agents via MCP, allowing seamless interaction and control. Here's how to get started with this powerful tool.

✨ Features

• Bidirectional Communication: Connect AI agents to Rhino via a socket-based server.
• Object Manipulation: Create, modify, and delete 3D objects in Rhino.
• Document Inspection: Retrieve detailed information about the current Rhino document.
• Script Execution: Execute Python scripts in Rhino (experimental and may not work every time).
• Script Documentation Retrieval: Obtain documentation for specific RhinoScript Python functions.
• Object Selection: Select objects based on filters (e.g., name, color, category) using "AND" or "OR" logic.
• Layer Management: Get or set the current layer, create new layers, or delete existing ones.

⚠️ Important Note

Currently, this tool only supports creating simple geometries as shown in the examples. More geometric types will be added in the future. Supported objects include points, lines, polylines, curves, boxes, and cylinders.

📦 Installation

1. Download and Install the RhinoMCP Plugin

# Enter the following command in the Rhino console
InstallPlugin("RhinoMCP.rhpkg")

2. Start the MCP Server

cd rhino_mcp_server
uv build
uv publish

💻 Usage Examples

Connecting an AI Agent

1. Launch the MCP plugin in Rhino.
2. Configure the plugin to listen on a specified port.
3. Connect the AI agent to that port.

Sending Commands

{
  "type": "create_box",
  "params": {
    "position": [0, 0, 0],
    "size": [1, 1, 1]
  }
}

Receiving Responses

{
  "status": "success",
  "result": "Box created successfully"
}

🔧 Technical Details

Communication Protocol

• Commands: Sent as JSON objects containing a type and optional params.
• Responses: JSON objects with status, result, or message.

Limitations and Security Precautions

• get_document_info retrieves a maximum of 30 objects to prevent Claude from being overloaded with large amounts of data.
• Complex operations may need to be broken down into smaller steps.

Tool Building and Publishing

Publishing the Server

cd rhino_mcp_server
uv build
uv publish

Publishing the Plugin

1. Build the tool in release mode.
2. Copy the "manifest.yml" file to the "bin/Release" folder.
3. Run yak build in the Release folder.
4. Run yak push rhino_mcp_plugin_xxxx.yak to publish the plugin.

Contribution

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

Disclaimer

This is a third-party integration not developed by Mcneel. It is created by Jingcheng Chen.

Star History