Jupyter MCP Server

Example Use Cases

• 1

  Inspect, edit, and execute cells inside a Jupyter notebook programmatically (read/write cell source, run cells, and capture outputs).

• 2

  Manage multiple notebooks and kernels remotely (list notebooks/kernels, switch or assign kernels, restart or unuse notebooks).

• 3

  Multimodal interaction with notebooks (handle images, plots, and rich outputs and use JupyterLab integration for automatic notebook opening).

Description

Jupyter MCP Server provides real-time control of Jupyter notebooks, including listing and managing notebooks and kernels, reading and editing cell source and outputs, and executing cells with multimodal support. It integrates with MCP clients (e.g., Claude Desktop, Cursor) and offers JupyterLab-specific features such as automatic notebook opening and native run-all commands. The server is designed for multi-notebook workflows, robust execution with feedback on failures, and supports images, plots, and other output types for richer agent interactions. It is developed and maintained by Datalayer with documentation and deployment options (uv/docker) available.

Capabilities(16 total)