Creativity · MCP — server

MCP Jupyter Notebook Server

The MCP Jupyter Notebook Server is a community-maintained Model Context Protocol server that connects to a Jupyter server or local kernel. It lets an LLM client list notebooks, create and edit cells, execute Python, and inspect rich outputs — an ideal pattern for data-analysis copilots that reason over execution results.

MCP facts

Kind
server
Ecosystem
anthropic-mcp
Language
Python
Transports
stdio

Capabilities

  • Tools: list_notebooks, read_notebook, insert_cell, execute_cell, get_cell_output
  • Resources: jupyter://notebook/{path} and jupyter://kernel/{id} URIs
  • Auth: JUPYTER_URL + JUPYTER_TOKEN env vars

Install

pip install mcp-server-jupyter && python -m mcp_server_jupyter

Configuration

{
  "mcpServers": {
    "jupyter": {
      "command": "python",
      "args": ["-m", "mcp_server_jupyter"],
      "env": {
        "JUPYTER_URL": "http://localhost:8888",
        "JUPYTER_TOKEN": "..."
      }
    }
  }
}

Frequently asked questions

Is this server official?

No — it is community-maintained. Jupyter has not released a first-party MCP server, but multiple quality implementations exist.

Can it run untrusted code?

Yes — execute_cell runs arbitrary Python in your kernel. Use a containerized Jupyter server for any code you did not write.

Does it support JupyterLab?

Yes, because JupyterLab uses the same Jupyter Server API. Most community servers work against JupyterLab, classic Notebook, and jupyter-server directly.

Sources

  1. Jupyter Server REST API — accessed 2026-04-20
  2. MCP community servers — accessed 2026-04-20

Related