PRIMS is a lightweight, open-source Model Context Protocol (MCP) server for running LLM agent-controlled Python code in secure, throw-away sandboxes. It enables reproducible, isolated code execution environments, installing dependencies and mounting files as needed, then destroying the environment after each use. Additional tools allow file inspection, previewing, downloading, and artifact persistence, making it ideal for developers, researchers, and AI agent integrators needing a code execution backend with strong isolation.
Visit Prims's official website for product details and getting started.
Comprehensive API reference and usage instructions for PRIMS.
Community-driven section for reporting bugs, requesting features, and discussing improvements.
Practical examples of how to use PRIMS in different scenarios.