jupyter-mcp-server

Why we included this project

Jupyter-mcp-server exposes a running Jupyter deployment as a set of tools that AI assistants can drive over the Model Context Protocol. Rather than reasoning about notebooks from a distance, an agent using an MCP client like Claude Desktop or Cursor can connect to your Jupyter server, browse the kernels and files it has, read and edit cell sources, and run code in a live kernel or an external sandbox such as Google Colab or Modal, with output and images streamed back into the conversation. That matters if you want your coding assistant to run Python inside the notebook workflow you already have instead of in a throwaway environment. Teams on JupyterHub or plain local installs can have assistants work on notebooks in place, and the built-in OpenTelemetry tracing gives you visibility into what the agent is doing. It is a focused bridge between the agent ecosystem and a data tool many teams already run, not a full product on its own.