Jupyter MCP Server

🚀 Jupyter MCP Server

Jupyter MCP Server is a Model Context Protocol (MCP) server implementation that provides interactive capabilities with 📓 Jupyter notebooks, which can run in any JupyterLab (including your local JupyterLab on 💻).

🚀 Quick Start

Start JupyterLab

Ensure the following packages are installed. The collaboration package is required because, thanks to Jupyter Real-Time Collaboration, it can display the modifications made to the notebook.

pip install jupyterlab jupyter-collaboration ipykernel
pip uninstall -y pycrdt datalayer_pycrdt
pip install datalayer_pycrdt

Then start JupyterLab using the following command:

jupyter lab --port 8888 --IdentityProvider.token MY_TOKEN --ip 0.0.0.0

You can also run make jupyterlab.

⚠️ Important Note

The --ip is set to 0.0.0.0 to allow the MCP server to access your local JupyterLab when running in a Docker container.

Use with Claude Desktop

Claude Desktop can be downloaded from this page for macOS and Windows.

For Linux, we successfully used the nix-based UNOFFICIAL build script

# ⚠️ UNOFFICIAL
# You can also run `make claude-linux`
NIXPKGS_ALLOW_UNFREE=1 nix run github:k3d3/claude-desktop-linux-flake \
  --impure \
  --extra-experimental-features flakes \
  --extra-experimental-features nix-command

To use it with Claude Desktop, add the following to your claude_desktop_config.json (for more details, see MCP official documentation website).

⚠️ Important Note

The value of SERVER_URL depends on where you run JupyterLab. If it's running locally, it should be set to http://localhost:8888; if it's through a cloud server, you need the corresponding URL and port.

Configuration Example

For macOS or Windows:

cat $HOME/.config/Claude/desktop_config/claude_desktop_config.json

The output might look like this:

{
  "mcp_server": {
    "add_execute_code_cell": {
      "description": "Add and execute a code cell in a Jupyter notebook",
      "parameters": [
        {
          "name": "cell_content",
          "type": "string",
          "description": "The code to execute"
        }
      ],
      "returns": "Cell output"
    },
    "add_markdown_cell": {
      "description": "Add a Markdown cell in a Jupyter notebook",
      "parameters": [
        {
          "name": "cell_content",
          "type": "string"
        }
      ]
    }
  }
}