RhinoMcpServer is an MCP server implementation intended to enable Claude to create and manipulate Rhino 3D objects via the MCP protocol. The project combines Python-based MCP server components with a C# Rhino plugin to bridge between the AI interface and Rhino. The server exposes a FastMCP-based set of tools for modeling operations and communicates through a combined MCP server entry point (combined_mcp_server.py) or through the daemon/socket-based alternatives as described in the launcher script. To get started, run the unified launcher and select the desired mode, then connect Claude Desktop (or your client) to the server. The available tools include geometry operations like creating spheres, boxes, and cylinders, as well as scene management utilities such as getting scene info, clearing the scene, and creating layers. The system is designed so that Python components handle MCP logic and tool methods, while a C# plugin handles Rhino-specific interactions and passes results back across the MCP channel.

Prerequisites:

• Python 3.8+ installed on your system
• Git to clone the repository
• Optional: Rhino 7+ with the RhinoCommon SDK if you plan to test Rhino integration locally

Installation steps:

1. Clone the repository git clone <repository-url> cd rhinoMcpServer

2. Create and activate a Python virtual environment (recommended) python -m venv .venv source .venv/bin/activate # On Windows use: .venv\Scripts\activate

3. Install Python dependencies (if a requirements file is provided) pip install -r requirements.txt # if the file exists

   If there is no requirements file, install any explicit dependencies mentioned in docs

4. Prepare server launcher (if using provided scripts) chmod +x server_launcher.sh # Unix-like systems

5. Run the server ./server_launcher.sh combined # or: python combined_mcp_server.py

6. Connect your MCP client (e.g., Claude Desktop) to the running server using the configured endpoint.

Notes:

• The project is marked as UNDER CONSTRUCTION, so some components may be unstable. Ensure Rhino integration components are properly compiled and reachable if testing Rhino-specific functionality.

Tips and caveats:

• This MCP server architecture pairs Python MCP components with a C# Rhino plugin. If Rhino integration fails, verify the Rhino connection layer (RhinoMcpPlugin.cs and related files) and ensure the Rhino plugin is loaded by Rhino.
• The launcher supports multiple modes (combined, standalone, daemon, socket-proxy, direct, logs, monitor, errors). Use server_launcher.sh help to inspect options and cleanup stale processes.
• Logging is centralized; check logs under the logs/ directory for component-specific messages. Use the launcher commands to view or monitor logs in real time.
• Since the project is under active development, expect breaking changes and evolving tool APIs. Keep the codebase synced with the latest docs and code_architecture.md for guidance.
• If you encounter connection issues, ensure there are no conflicting server instances and that the appropriate ports/sockets are accessible by the client.