jupyter-mcp-tools

What it is and what it does

jupyter-mcp-tools is a JupyterLab server extension that bridges JupyterLab's command system with the Model Context Protocol (MCP), a standardized interface for AI agents and external tools. It dynamically registers all available JupyterLab commands—notebook operations, file management, kernel control, console commands, and UI toggles—as MCP tools that can be invoked remotely via WebSocket.

The extension operates in two modes: local mode executes commands directly within the JupyterLab frontend, while remote mode transmits commands via WebSocket to the backend, enabling external MCP clients and AI agents to control JupyterLab from outside the environment. It depends on aiohttp for async HTTP/WebSocket handling and jupyter-server for the backend infrastructure, and is designed to work with jupyter-mcp-server for full MCP integration.

Use it for:

• Enable AI agents to open notebooks, execute cells, and manage kernel state in JupyterLab through a standardized MCP interface.
• Automate notebook workflows (append cells, run all cells, manage cell selection) triggered by external systems or chatbots.
• Provide remote file browser and document management operations (create, delete, rename, open files) to external MCP clients.
• Allow external tools to inject code into consoles, restart kernels, and manage terminal sessions without direct UI access.
• Integrate JupyterLab command execution into AI agent orchestration platforms that speak the MCP protocol.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Exposes JupyterLab commands as MCP (Model Context Protocol) tools, enabling external AI agents and clients to trigger notebook operations, file management, and kernel control through a standardized protocol.

Yes, if you need to expose JupyterLab commands to external MCP clients or AI agents. The package is actively maintained, has no known vulnerabilities, and carries a permissive BSD license. Install friction is low. Primary consideration: verify whether the frontend NPM package must be installed separately and whether jupyter-mcp-server is required for your use case.

Install

jupyter-mcp-tools on PyPI

pip install jupyter-mcp-tools

uv add jupyter-mcp-tools

poetry add jupyter-mcp-tools

Installing jupyter-mcp-tools

Before you install

Low friction: pure Python wheel with only aiohttp and jupyter-server as runtime dependencies. Active maintenance with a release 4 days old; repository is not archived and supports Python 3.9 through 3.13.

License in practice

BSD 3-Clause permissive license allows commercial and private use with minimal restrictions; you may use, modify, and distribute the software provided you retain the license notice.

Quickstart

pip install jupyter-mcp-tools

# Then enable in JupyterLab and connect via MCP client:
import aiohttp
from jupyter_mcp_tools import MCP_TOOLS

# MCP tools are automatically registered and exposed via WebSocket

Requires a running Jupyter Server instance and a corresponding MCP client to connect; the frontend NPM package may need separate installation.

Verify before relying

• Whether the extension requires a corresponding frontend NPM package (@datalayer/jupyter-mcp-tools) to be installed separately for full functionality.
• Whether jupyter-mcp-server is a required peer dependency or an optional integration point.
• Performance characteristics when exposing a large number of JupyterLab commands as MCP tools.

Package facts

Evidence: jupyter_mcp_tools-0.1.7-py3-none-any.whl

Keywords: ai, jupyter, jupyterlab, jupyterlab-extension, mcp

Tags

More Application Frameworks packages

Further reading