ipybox

ipybox is a lightweight and secure Python code execution sandbox based on IPython and Docker. You can run it locally on your computer or remotely in an environment of your choice. ipybox is designed for AI agents that need to execute code safely e.g. for data analytics use cases or executing code actions like in freeact agents.

Features

• Secure code execution inside Docker containers
• Restrict network access with a configurable firewall
• Stateful code execution with IPython kernels
• Stream code execution output as it is generated
• Install Python packages at build time or runtime
• Return plots generated with visualization libraries
• Exposes an MCP server interface for AI agent integration
• Invocation of MCP servers via generated client code
• Flexible deployment options, local or remote
• asyncio API for managing the execution environment

Documentation

• User Guide
• API Docs

MCP server

ipybox can be installed as MCP server.

{
  "mcpServers": {
    "ipybox": {
      "command": "uvx",
      "args": ["ipybox", "mcp"]
    }
  }
}

Python example

Install ipybox Python package:

pip install ipybox

Execute code in an ipybox container:

import asyncio
from ipybox import ExecutionClient, ExecutionContainer

async def main():
    async with ExecutionContainer(tag="ghcr.io/gradion-ai/ipybox") as container:
        async with ExecutionClient(port=container.executor_port) as client:
            result = await client.execute("print('Hello, world!')")
            print(f"Output: {result.text}")

if __name__ == "__main__":
    asyncio.run(main())