python-mcp-server

python-mcp-server

3.3

This MCP server runs a Python environment to execute code generated by an LLM, useful for web scraping.

The Sample MCP Server in Go is designed to facilitate the execution of code generated by a Language Learning Model (LLM) within a Python environment. This server is particularly useful for tasks such as web scraping. The project is a modified version of a demo available on YouTube and utilizes Podman as the container engine instead of Docker. The server is built using Go 1.23+ and can be tested using the 'mcphost' tool. Configuration involves setting up a JSON file to define the command for the MCP server, and the server can be run using Claude-3.5 Sonnet as the default LLM, which does not require a local GPU. The server allows users to prompt requests for web scraping and other tasks, showcasing the integration of LLMs with practical applications.

Features

  • Python Environment: Executes code generated by LLMs for tasks like web scraping.
  • Podman Integration: Uses Podman as the container engine instead of Docker.
  • Go Language: Built using Go 1.23+ for efficient performance.
  • LLM Compatibility: Works with Claude-3.5 Sonnet as the default LLM.
  • Configurable: Easily configured through a JSON file for command setup.

Usage with Different Platforms

Build Project

shell
$ go mod tidy && go install

Configure MCP Server


{
  "mcpServers": {
      "python-repl": {
          "command": "python-mcp-server"
      }
  }
}

Run mcphost

shell
$ export ANTHROPIC_API_KEY='your-api-key'
$ mcphost