RhinoMCP - Rhino Model Context Protocol Integration

RhinoMCP connects Rhino to AI agents through the Model Context Protocol (MCP), allowing AI agents to directly interact with and control Rhino. This integration enables prompt assisted 3D modeling in Rhino 3D.

Features

• Two-way communication: Connect AI agents to Rhino through a socket-based server
• Object manipulation: Create, modify, and delete 3D objects in Rhino
• Document inspection: Get detailed information about the current Rhino document
• Script execution: Execute Rhinos python scripts in Rhino (experimental, may not work every time)
• Get Script Documentation: Get the documentation of a specific RhinoScript python function
• Object selection: Select objects based on filters, e.g. name, color, category, etc. with "and" or "or" logic
• Set/Create/Delete Layers: Get or set the current layer, create new layers, or delete layers

[!NOTE]
So far the tool only supports creating primitive objects for proof of concept. More geometries will be added in the future. Supported objects: Point, Line, Polyline, Circle, Arc, Ellipse, Curve, Box, Sphere, Cone, Cylinder, Surface (from points)

Demo

Demo 1

This demo shows how AI can interact with Rhino in two directions. Click the image below to watch the video.

Demo 2

This demo shows how to ask AI to create custom scripts and execute them in Rhino. Click the image below to watch the video.

Tutorial

Thanks to Nate. He has created a showcase and installation tutorial for this tool.

Components

The system consists of two main components:

1. MCP Server (src/rhino_mcp_server/server.py): A Python server that implements the Model Context Protocol and connects to the Rhino plugin
2. Rhino Plugin (src/rhino_mcp_plugin): A Rhino plugin that creates a socket server within Rhino to receive and execute commands

Installation

Prerequisites

• Rhino 8 (Windows or macOS). The plugin targets net8.0 against RhinoCommon 8 and is not built for Rhino 7.
• Python 3.10 or newer
• uv package manager

⚠️ Only run one instance of the MCP server (either on Cursor or Claude Desktop), not both

Installing the Rhino Plugin

1. Go to Tools > Package Manager
2. Search for rhinomcp
3. Click Install

Install uv

If you're on Mac, please install uv as

brew install uv

On Windows

powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

⚠️ Do not proceed before installing UV

Config file

{
  "mcpServers": {
    "rhino": {
      "command": "uvx",
      "args": ["rhinomcp"]
    }
  }
}

or if you want your rhinomcp process to be stopped with your AI Agent (mac & linux):

{
  "mcpServers": {
      "rhino": {
        "command": "sh",
        "args": ["-c", "killall rhinomcp 2>/dev/null; uvx rhinomcp"]
      }
    }
}

and for Windows:

{
  "mcpServers": {
      "rhino": {
        "command": "cmd",
        "args": ["/c", "taskkill /F /IM rhinomcp.exe 2>nul & uvx rhinomcp"]
      }
    }
}

Claude for Desktop Integration

Go to Claude > Settings > Developer > Edit Config > claude_desktop_config.json to include the above config file.

Cursor integration

Make sure your cursor is up to date.

Create a folder .cursor in your project root.

Create a file mcp.json in the .cursor folder and include the above config file:

Go to Cursor Settings > MCP and check if it's enabled.

Security posture

The Python MCP server and the Rhino plugin talk over an unauthenticated TCP loopback link (default 127.0.0.1:1999). Tools like run_command, execute_rhinoscript_python_code, and execute_rhinocommon_csharp_code hand the model a fully open execution su

... View full README on GitHub