Skip to main content
Glama
mcp-pyodide
by yonaka15

mcp-pyodide

A Pyodide server implementation for the Model Context Protocol (MCP). This server enables Large Language Models (LLMs) to execute Python code through the MCP interface.

Features

  • Python code execution capability for LLMs using Pyodide

  • MCP compliant server implementation

  • Support for both stdio and SSE transport modes

  • Robust implementation written in TypeScript

  • Available as a command-line tool

Related MCP server: Sandbox MCP

Installation

npm install mcp-pyodide

Usage

As a Server

import { runServer } from "mcp-pyodide";

// Start the server
runServer().catch((error: unknown) => {
  console.error("Error starting server:", error);
  process.exit(1);
});

As a Command-line Tool

Start in stdio mode (default):

mcp-pyodide

Start in SSE mode:

mcp-pyodide --sse

SSE Mode

When running in SSE mode, the server provides the following endpoints:

  • SSE Connection: http://localhost:3020/sse

  • Message Handler: http://localhost:3020/messages

Example client connection:

const eventSource = new EventSource("http://localhost:3020/sse");
eventSource.onmessage = (event) => {
  console.log("Received:", JSON.parse(event.data));
};

Project Structure

mcp-pyodide/
├── src/
│   ├── formatters/    # Data formatting handlers
│   ├── handlers/      # Request handlers
│   ├── lib/          # Library code
│   ├── tools/        # Utility tools
│   ├── utils/        # Utility functions
│   └── index.ts      # Main entry point
├── build/            # Build artifacts
├── pyodide-packages/ # Pyodide-related packages
└── package.json

Dependencies

  • @modelcontextprotocol/sdk: MCP SDK (^1.4.0)

  • pyodide: Python runtime environment (^0.27.1)

  • arktype: Type validation library (^2.0.1)

  • express: Web framework for SSE support

  • cors: CORS middleware for SSE support

Development

Requirements

  • Node.js 18 or higher

  • npm 9 or higher

Setup

# Clone the repository
git clone <repository-url>

# Install dependencies
npm install

# Build
npm run build

Scripts

  • npm run build: Compile TypeScript and set execution permissions

  • npm start: Run server in stdio mode

  • npm run start:sse: Run server in SSE mode

Environment Variables

  • PYODIDE_CACHE_DIR: Directory for Pyodide cache (default: "./cache")

  • PYODIDE_DATA_DIR: Directory for mounted data (default: "./data")

  • PORT: Port for SSE server (default: 3020)

License

MIT

Contributing

  1. Fork the repository

  2. Create a feature branch (git checkout -b feature/amazing-feature)

  3. Commit your changes (git commit -am 'Add some amazing feature')

  4. Push to the branch (git push origin feature/amazing-feature)

  5. Create a Pull Request

Important Notes

  • This project is under development, and the API may change

  • Thoroughly test before using in production

  • Exercise caution when executing untrusted code for security reasons

  • When using SSE mode, ensure proper CORS configuration if needed

Support

Please use the Issue tracker for problems and questions.

Available Tools

5 tools
pyodide_executeA

Execute Python code using Pyodide with output capture. When generating images, they will be automatically saved to the output directory instead of being displayed. Images can be accessed from the saved file paths that will be included in the output.

Inputs: code (required), timeout.

pyodide_get-mount-pointsB

List mounted directories

Inputs: no named parameters.

pyodide_install-packagesC

Install Python packages using Pyodide. Multiple packages can be specified using space-separated format.

Inputs: package (required).

pyodide_list-mounted-directoryB

List contents of a mounted directory

Inputs: mountName (required).

pyodide_read-imageC

Read an image from a mounted directory

Inputs: mountName (required), imagePath (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3.4/5.0

    Scored across 5 tools

    Disambiguation4/5

    Most tools have distinct purposes: execute code, list mount points, install packages, list directory contents, and read images. However, pyodide_get-mount-points and pyodide_list-mounted-directory could cause minor confusion as both relate to directory listing, but their scopes differ (mount points vs. specific directory contents).

    Naming Consistency5/5

    All tools follow a consistent pyodide_verb-noun naming pattern with hyphens for multi-word verbs or nouns. This uniformity makes the tool set predictable and easy to understand, with no mixing of conventions.

    Tool Count4/5

    With 5 tools, the count is reasonable for a Pyodide execution server, covering core operations like code execution, package management, and file handling. It might benefit from additional tools for writing files or managing environments, but it's well-scoped for basic functionality.

    Completeness3/5

    The tools cover key areas such as code execution, package installation, and file reading/listing, but there are notable gaps. For example, there's no tool for writing or saving files (beyond automatic image saving), deleting files, or managing Python environments more comprehensively, which could limit agent workflows.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      A Model Context Protocol (MCP) server that enables LLMs to run ANY code safely in isolated Docker containers.
      124
      MIT
    • A
      license
      B
      quality
      D
      maintenance
      A secure Model Context Protocol server that allows AI assistants and LLM applications to safely execute Python and JavaScript code snippets in containerized environments.
      2
      206
      MIT
    • A
      license
      A
      quality
      D
      maintenance
      A production-grade MCP server providing a persistent Python REPL with multi-session support, sandboxing, and timeout protection, enabling LLM agents to execute Python code across multiple turns with variables that persist between calls.
      12
      31 PyPI
      1
      MIT