Python Notebook MCP

Example Use Cases

• 1

  Discover and list all .ipynb files inside a user's project workspace to choose or present notebooks.

• 2

  Read, create, and modify notebook content and cells programmatically (add/edit cells, read cell outputs, read full notebook structure).

• 3

  Integrate with MCP-compatible clients (e.g., Cursor or Claude Desktop) and run in a controlled local virtual environment or via fastmcp.

Description

This MCP server exposes tools for listing, reading, creating, and editing Jupyter notebooks and individual cells, as well as reading cell outputs. It is designed to run locally in a virtual environment or via fastmcp for Claude Desktop integration, enabling safe, direct access to a user's workspace. The server requires an explicit workspace initialization (absolute path) and aims to simplify integration with MCP-compatible clients like Cursor and Claude Desktop. The project is open-source (MIT) and includes install scripts for common platforms.