Datalayer, a company building collaborative Jupyter infrastructure, publishes and maintains this server.

An MCP server that connects an AI agent to a running Jupyter server so it can read, edit, and execute notebook cells in real time.

Runs locally or against a remote JupyterLab or Jupyter Server instance over its REST and kernel APIs.

Published by Datalayer in 2024 and actively developed through 2025 alongside the wider notebook MCP ecosystem.

Lets an agent iterate on data analysis or ML code by running cells directly instead of only editing static files.

Requires an MCP client and a running Jupyter server URL plus its authentication token.

Community vendor maintained by Datalayer, source available under a permissive license, can execute code and modify notebook state so it can change data and files, runs against a local or remote Jupyter server.