ipybox is a lightweight and secure Python code execution sandbox based on IPython and Docker. It can run locally on your computer or remotely in an environment of your choice. It is designed for AI agents that need to execute Python code securely, for tasks such as data analysis or executing code actions. The sandbox runs inside isolated Docker containers with a stateful IPython kernel, and offers a configurable firewall to restrict network access. It streams code execution output as it is generated and can install Python packages at build time or runtime. It can return plots generated with visualization libraries. The project exposes an MCP server interface for AI agent integration, and provides generated client code to invoke MCP servers. Deployment is flexible, supporting local or remote setups, and an asyncio API is available for managing the execution environment and coordinating concurrent containers. Together, ipybox enables secure, reproducible Python code execution within an MCP workflow.