Python Sandbox

Provides secure Python code execution in containerized sandboxes with automatic matplotlib figure persistence and comprehensive data science libraries for interactive analysis and machine learning experimentation.

This MCP server provides secure Python code execution capabilities through a containerized sandbox environment optimized for data science and machine learning workflows. Built by aamir-gmail specifically for LM Studio integration, it features automatic matplotlib figure persistence, per-execution isolation with unique run directories, and dual deployment modes supporting both LM Studio's native MCP chat interface and a Gradio web UI through the developer API. The implementation includes a comprehensive data science stack (numpy, pandas, scikit-learn, matplotlib, seaborn, xgboost) with automatic artifact serving, making it valuable for interactive data analysis, visualization generation, and machine learning experimentation where users need reliable code execution with persistent file outputs accessible via web URLs.