Model Context Protocol (MCP) implementation for Jupyter notebooks to overcome limitation with Cursor
Connect Cursor to Claude, Cursor or any other MCP client and it stops being a tab you switch to. Model Context Protocol (MCP) implementation for Jupyter notebooks to overcome limitation with Cursor. The cursor mcp server is what makes that connection.
This directory contains a Model Context Protocol (MCP) server designed to allow AI agents within Cursor to interact with Jupyter Notebook (.ipynb) files. It was created to overcome a limitation with Cursor. As of version 0.50.5, in Agent mode, the model could not edit notebooks or notebook cells in response to dialog in the AI chat pane. This provides the agent with a suite of MCP tools that allow direct notebook cell manipulation.
Installation goes through your MCP client rather than a global install: point it at cursor-notebook-mcp on PyPI and it is fetched when the client starts. The copy-paste blocks for Claude Desktop, Claude Code and Cursor are further down this page.
The toolset is worth reading before you wire it up, because it tells you what the integration is really for:
notebook_create — Creates a new, empty notebook filenotebook_delete — Deletes an existing notebook filenotebook_rename — Renames/moves a notebook file from one path to anothernotebook_read — Reads an entire notebook and returns its structure as a dictionarynotebook_read_cell — Reads the source content of a specific cellnotebook_add_cell — Adds a new code or markdown cell after a specified indexnotebook_edit_cell — Replaces the source content of a specific cellnotebook_delete_cell — Deletes a specific cellnotebook_change_cell_type — Changes a cell's type (code, markdown, or raw)notebook_duplicate_cell — Duplicates a cell multiple times (default: once)notebook_get_cell_count — Returns the total number of cellsnotebook_read_metadata — Reads the top-level notebook metadataThis project has both Python package dependencies and potentially external system dependencies for full functionality.
This sits in the developer tooling group, where several servers overlap in what they claim to do but differ sharply once you actually set them up. Cursor's toolset — notebook_create, notebook_delete, notebook_rename and 11 more — is a fair guide to whether it matches your workflow.
This entry was verified against Cursor's own documentation before publication; SyncDev keeps the directory reviewed rather than auto-generated.
| Tool | What it does |
|---|---|
| notebook_create | Creates a new, empty notebook file. |
| notebook_delete | Deletes an existing notebook file. |
| notebook_rename | Renames/moves a notebook file from one path to another. |
| notebook_read | Reads an entire notebook and returns its structure as a dictionary. |
| notebook_read_cell | Reads the source content of a specific cell. |
| notebook_add_cell | Adds a new code or markdown cell after a specified index. |
| notebook_edit_cell | Replaces the source content of a specific cell. |
| notebook_delete_cell | Deletes a specific cell. |
| notebook_change_cell_type | Changes a cell's type (code, markdown, or raw). |
| notebook_duplicate_cell | Duplicates a cell multiple times (default: once). |
| notebook_get_cell_count | Returns the total number of cells. |
| notebook_read_metadata | Reads the top-level notebook metadata. |
| notebook_edit_metadata | Updates the top-level notebook metadata. |
| notebook_read_cell_metadata | Reads the metadata of a specific cell. |
{
"mcpServers": {
"cursor-notebook": {
"command": "uvx",
"args": ["cursor-notebook-mcp"]
}
}
}Add to claude_desktop_config.json, then restart Claude Desktop.
This project has both Python package dependencies and potentially external system dependencies for full functionality.
Kill hallucinated APIs — version-accurate, up-to-date library documentation injected straight into context.
Microsoft's official browser automation server — drive a real browser through the accessibility tree, no screenshots needed.
GitHub's official server — repos, issues, pull requests, Actions and code security, straight from your assistant.
Issue tracking at the speed of conversation — Linear's official hosted server with OAuth and zero install.
Local repository surgery — status, diffs, commits, branches and history for any repo on disk.
Timezone sanity for AI — current time anywhere and correct conversions, without the model doing date math.