Jupyter MCP Server

@MCP-Mirror

リポジトリ

基本情報

カテゴリ

データと分析

トランスポート

stdio

作者

MCP-Mirror

Jupyter MCP Server について

Mirror of

#jupyter #mcp-server #data-science

Jupyter MCP Server の設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "datalayer_jupyter-mcp-server": {
      "command": "docker",
      "args": [
        "build",
        "-t",
        "datalayer/jupyter-mcp-server",
        "."
      ]
    }
  }
}

Jupyter MCP Server のツール

`cell_content`(string): Code to be executed

Success message

`cell_content`(string): Markdown content

Success message

概要

What is Jupyter MCP Server?

Jupyter MCP Server is a Model Context Protocol (MCP) server that enables interaction with Jupyter notebooks running in a local JupyterLab instance. Designed for AI agents and developers, it uses Jupyter Real Time Collaboration (RTC) to reflect notebook modifications live.

How to use Jupyter MCP Server?

Start JupyterLab with token authentication (jupyter lab --port 8888 --IdentityProvider.token MY_TOKEN --ip 0.0.0.0) and install the required packages (jupyterlab, jupyter-collaboration, ipykernel). Then configure the MCP client (e.g., Claude Desktop) to run the Docker image datalayer/jupyter-mcp-server:latest, setting environment variables SERVER_URL, TOKEN, and NOTEBOOK_PATH to match your JupyterLab session.

Key features of Jupyter MCP Server

Connect AI agents to live Jupyter notebooks
Add and execute code cells remotely
Insert Markdown cells into notebooks
Uses Docker for easy deployment
Works with JupyterLab’s Real Time Collaboration

Use cases of Jupyter MCP Server

Enable a language model to run Python code in a Jupyter notebook
Automatically generate and insert analysis cells with results
Combine AI-generated explanations with computed outputs in a single notebook

よくある質問

What tools does Jupyter MCP Server offer?

Two tools: `add_execute_code_cell` (executes Python code and returns success) and `add_markdown_cell` (adds Markdown content). Both take a `cell_content` string.

What are the runtime dependencies?

A local JupyterLab with `jupyter-collaboration` and `ipykernel` installed, plus Docker to run the MCP server container.

How do I authenticate to my JupyterLab instance?

You must set a token via `--IdentityProvider.token MY_TOKEN` when starting JupyterLab, then provide the same token in the `TOKEN` environment variable in the MCP client configuration.

How does the server connect to a running JupyterLab?

The `SERVER_URL` environment variable (e.g., `http://localhost:8888`) tells the Docker container where JupyterLab is listening. On Linux, the `--network=host` Docker flag is used; on macOS/Windows, the container accesses the host via `host.docker.internal`.

Can I install Jupyter MCP Server without Docker?

The README only shows distribution via Docker or Smithery. There is no pip-based installation described.

「データと分析」の他のコンテンツ

PropertyRadar

Research U.S. markets and prospects conversationally using PropertyRadar's property and people data. Resolve an address, inspect Property details, transactions, comparable sales, ownership and contact information, discov

CovaSyn Chemistry Tools

CovaSyn GmbH

Chemistry, pharma and biotech tools for AI assistants such as Claude, ChatGPT and Cursor. Deterministic calculations instead of guesses: NMR, MS, IR and UV spectra analysis, ADMET and ICH M7 mutagenicity assessment, stab

Staats

Mr Violets

Cookieless web analytics your coding agent reads and remembers. Traffic, referrers and funnels over MCP, plus deploy-tagged before/after comparisons, stored site context, and proactive anomaly surfacing. Remote server, n

Placestack

Parcels, owners, assessments, sales, zoning rules, building permits and zoning cases for any U.S. address.

Ultimate Web Scraper

Extract products, prices, images and reviews from any page in one click — or connect your AI agent and let it do the scraping for you. Works great for: - Ecommerce: extract every product from a Shopify, WooCommerce or