Jupyter MCP Server
Let Claude read, edit and run cells in your live Jupyter notebooks.
Community, verifiedby DatalayerRuns on your computerChecked October 18, 2026
This open-source MCP server connects Claude to a running JupyterLab or JupyterHub, so it can open notebooks, insert and edit cells, execute code and read the outputs, including charts. It is maintained by Datalayer and works with any Jupyter setup that has real-time collaboration turned on.
Who it is for
- Data scientists
- Analysts who work in notebooks
- Students learning Python for data analysis
Prompts to try
Once it is connected, ask in plain words. Claude decides when to use the server and asks before it changes anything important.
Open sales_analysis.ipynb, load the CSV in the first cell and plot monthly revenue.Find the cell that throws an error in this notebook, fix it and rerun everything below it.Add a section that compares churn by plan type and explains the result in a markdown cell.How to connect it
In Claude Code
claude mcp add --transport stdio jupyter --env JUPYTER_URL=http://localhost:8888 --env JUPYTER_TOKEN=YOUR_JUPYTER_TOKEN --env ALLOW_IMG_OUTPUT=true -- uvx jupyter-mcp-server@latest
In the Claude desktop app
Add this to claude_desktop_config.json, then restart the app:
{
"mcpServers": {
"jupyter": {
"command": "uvx",
"args": [
"jupyter-mcp-server@latest"
],
"env": {
"JUPYTER_URL": "http://localhost:8888",
"JUPYTER_TOKEN": "YOUR_JUPYTER_TOKEN",
"ALLOW_IMG_OUTPUT": "true"
}
}
}
}Good to know
Start JupyterLab with a token and jupyter-collaboration installed before connecting. Datalayer also hosts the server at https://mcp.datalayer.run/mcp for its own cloud notebooks.
Unlock the setup steps
Free. Leave your email once to open the setup steps for all 150 servers, and get the best new connectors each month.
How we checked it
Checked the datalayer/jupyter-mcp-server repo (BSD-3-Clause, about 1,300 stars, release v2.2.4 on 2026-10-11) and the jupyter-mcp-server package on PyPI; it is built by Datalayer, not by Project Jupyter.