mcp-pyodide

What is mcp-pyodide?

mcp-pyodide is a server implementation of the Model Context Protocol (MCP) built on Pyodide. It enables Large Language Models (LLMs) to execute Python code through the MCP interface, allowing AI assistants to run Python scripts and analyze results programmatically.

How to use mcp-pyodide?

Install the package with npm install mcp-pyodide. Use it programmatically by importing runServer, or as a command‑line tool by running mcp-pyodide for stdio mode or mcp-pyodide --sse for SSE mode. When running in SSE mode, the server exposes an SSE endpoint at http://localhost:3020/sse and a message handler at http://localhost:3020/messages.

Key features of mcp-pyodide

• Python code execution for LLMs using the Pyodide runtime
• MCP compliant server implementation
• Supports both stdio and SSE transport modes
• Written in TypeScript with robust error handling
• Available as a command‑line tool

Use cases of mcp-pyodide

• Allowing an LLM to run and evaluate Python code snippets
• Enabling MCP‑compatible clients to execute dynamic Python logic
• Providing a controlled Python sandbox within an AI workflow
• Testing and debugging Python scripts via an LLM interface
• Running data analysis or small‑scale computations in a serverless fashion

FAQ from mcp-pyodide

What are the system requirements for mcp-pyodide?

Node.js 18 or higher and npm 9 or higher are required.

What transport modes does mcp-pyodide support?

It supports stdio mode (default) and SSE (Server‑Sent Events) mode via the --sse flag.

How can I configure the server’s cache or data directories?

Use the environment variables PYODIDE_CACHE_DIR (default ./cache), PYODIDE_DATA_DIR (default ./data), and PORT (default 3020 for SSE mode).

Is mcp-pyodide safe for production use?

The README notes the project is under active development and the API may change. It advises thorough testing before production use and warns against executing untrusted code without proper security precautions.

How do I contribute to mcp-pyodide?

Fork the repository, create a feature branch, commit changes, and submit a pull request following the standard GitHub workflow.