Python notebook

What is Python notebook?

Python notebook is an MCP (Model Context Protocol) server that enables AI assistants (such as Cursor or Claude Desktop) to interact with Jupyter Notebook files (.ipynb) on your local machine. It is designed for developers and data scientists who want to use AI tools to programmatically manage notebooks.

How to use Python notebook?

To use Python notebook, ensure Python 3.10+ and uv are installed, then clone the repository and run the automated setup script (install_unix.sh on macOS/Linux or install_windows.ps1 on Windows) or manually create a virtual environment and install dependencies from requirements.txt. The server is started with uv run python server.py (for Cursor) or fastmcp install server.py --name "Jupyter Notebook MCP" (for Claude Desktop). The first action from your AI assistant must be to call initialize_workspace(directory="/full/absolute/path/to/project").

Key features of Python notebook

• Initialize workspace with an absolute directory path.
• List all .ipynb files in the workspace.
• Create new empty Jupyter notebooks.
• Read entire notebook structure and content.
• Add, read, and edit code or markdown cells.
• Read outputs from specific or all code cells.

Use cases of Python notebook

• Automatically create and populate new Jupyter notebooks from an AI prompt.
• Edit existing notebook cells without leaving the AI assistant interface.
• Inspect notebook contents and outputs for debugging or documentation.
• Enable AI-assisted data analysis workflows inside a local notebook environment.
• Manage multiple notebook projects through a centralized workspace configuration.

FAQ from Python notebook

What is the first action required after starting the server?

You must call the initialize_workspace tool with the full absolute path to a directory containing your notebooks. Relative paths or dots (.) are not accepted.

Can I edit existing cells in a notebook?

Yes, the edit_cell tool modifies the source content of an existing cell by its index, and add_cell adds new code or markdown cells at a specific index or the end.

What runtime dependencies does Python notebook require?

Python 3.10 or higher, the uv package manager, and optionally the fastmcp CLI (for Claude Desktop integration). All other dependencies are installed from requirements.txt.

Where are my notebook files stored?

Notebooks reside on your local filesystem in the workspace directory you provide during initialize_workspace. The server does not copy or move files.

What transport or authentication does Python notebook use?

The server uses standard MCP stdio transport for communication with the client. No authentication is mentioned in the README.