🪐✨ Jupyter MCP Tools

This extension is composed of a Python package named jupyter_mcp_tools for the server extension and a NPM package named @datalayer/jupyter-mcp-tools for the frontend extension.

It enables the use of JupyterLab commands as MCP tools.

This extension is used by jupyter-mcp-server to enable JupyterLab commands such as opening notebooks through MCP tools.

Execution Modes

The extension supports two execution modes for JupyterLab commands:

Local Mode

• Direct Execution: Commands are executed directly within the JupyterLab frontend using the built-in command registry
• No Network Required: Immediate execution without WebSocket communication
• Use Case: Testing commands and direct UI interaction within JupyterLab itself
• Implementation: Calls app.commands.execute() directly in the browser

Remote Mode

• WebSocket Communication: Commands are sent via WebSocket to the backend server extension
• External Access: Enables external MCP clients to execute JupyterLab commands remotely
• Use Case: Integration with AI agents and MCP clients like jupyter-mcp-server
• Implementation: Messages are transmitted through WebSocket protocol to backend handlers

The remote mode is what makes this extension valuable for MCP integration - it allows AI agents to trigger JupyterLab commands (like opening notebooks, executing cells, etc.) from outside the JupyterLab environment through a standardized protocol.

Available Tools

The jupyter-mcp-tools extension dynamically registers all available JupyterLab commands as MCP tools. The exact list of available tools depends on the installed JupyterLab extensions and the current context, but here are some commonly available tools :

Note: This extension automatically discovers and makes available all JupyterLab commands as MCP tools. Command IDs are converted from namespace:command format to namespace_command for MCP compatibility. The list below represents commonly available commands but is not exhaustive.

Core Notebook Commands

Tool ID	Description	Parameters
notebook_run-all-cells	Execute all cells in the current notebook sequentially	None
notebook_get-selected-cell	Get information about the currently selected cell	None
notebook_append-execute	Append a new cell at the end of notebook with source code and execute it	source (string), type (code/markdown/raw)
notebook_insert-cell-below	Insert a new cell below the current cell	activate (boolean)
notebook_insert-cell-above	Insert a new cell above the current cell	activate (boolean)
notebook_delete-cell	Delete the currently selected cell	None
notebook_cut-cell	Cut the currently selected cell	None
notebook_copy-cell	Copy the currently selected cell	None
notebook_paste-cell-below	Paste cell below the current cell	None
notebook_paste-cell-above	Paste cell above the current cell	None
notebook_select-next	Select the next cell	None
notebook_select-previous	Select the previous cell	None
notebook_extend-selection-below	Extend selection to cell below	None
notebook_extend-selection-above	Extend selection to cell above	None
notebook_move-cell-up	Move current cell up	None
notebook_move-cell-down	Move current cell down	None
notebook_split-cell	Split current cell at cursor position	None
notebook_merge-cell-above	Merge current cell with the one above	None
notebook_merge-cell-below	Merge current cell with the one below	None
notebook_run-cell	Execute the currently selected cell	None
notebook_run-cell-and-select-next	Execute cell and move to next	None
notebook_run-cell-and-insert-below	Execute cell and insert new cell below	None

Console Commands

Tool ID	Description	Parameters
console_create	Create a new console	activate (boolean), insertMode (string), path (string)
console_clear	Clear the console output	None
console_restart-kernel	Restart the console kernel	None
console_interrupt-kernel	Interrupt the console kernel	None
console_inject	Inject code into console	code (string), activate (boolean)

File/Document Management

Tool ID	Description	Parameters
docmanager_open	Open a document by path	path (string), factory (string), kernel (object)
docmanager_new-untitled	Create a new untitled document	type (string), path (string)
docmanager_save	Save the current document	None
docmanager_save-as	Save document with a new name	None
docmanager_rename	Rename the current document	newName (string)
docmanager_delete	Delete a document	path (string)
docmanager_duplicate	Duplicate a document	path (string)

File Browser Commands

Tool ID	Description	Parameters
filebrowser_go-to-path	Navigate to a specific path	path (string)
filebrowser_refresh	Refresh the file browser	None
filebrowser_toggle-hidden-files	Toggle showing hidden files	None
filebrowser_create-new-directory	Create a new directory	path (string)
filebrowser_upload	Upload files	path (string)
filebrowser_download	Download files	path (string)

Kernel Management

Tool ID	Description	Parameters
kernel_restart	Restart the current kernel	None
kernel_interrupt	Interrupt the current kernel	None
kernel_shutdown	Shutdown the current kernel	None
kernel_reconnect	Reconnect to kernel	None
kernel_change	Change to a different kernel	kernel (object)

UI/Layout Commands

Tool ID	Description	Parameters
application_toggle-left-area	Toggle the left sidebar	None
application_toggle-right-area	Toggle the right sidebar	None
application_toggle-presentation-mode	Toggle presentation mode	None
application_set-theme	Change the UI theme	theme (string)
mainmenu_open-edit	Open the Edit menu	None
mainmenu_open-file	Open the File menu	None
mainmenu_open-help	Open the Help menu	None

Cell Type Conversion

Tool ID	Description	Parameters
notebook_change-cell-to-code	Convert cell to code cell	None
notebook_change-cell-to-markdown	Convert cell to markdown cell	None
notebook_change-cell-to-raw	Convert cell to raw cell	None

Search and Replace

Tool ID	Description	Parameters
documentsearch_start	Start document search	query (string), caseSensitive (boolean), wholeWord (boolean), regex (boolean)
documentsearch_highlightNext	Highlight next search result	None
documentsearch_highlightPrevious	Highlight previous search result	None

Terminal Commands

Tool ID	Description	Parameters
terminal_create-new	Create a new terminal	cwd (string)
terminal_refresh	Refresh terminal	None

To see all available tools in your specific JupyterLab installation, you can open the MCP Tools panel in the left sidebar to browse available tools.

Additional Commands

This extension provides custom JupyterLab commands specifically designed for MCP integration:

notebook:append-execute

Appends a new cell at the end of the current notebook with the given source code and optionally executes it.

Parameters:

• source (string, required): The source code to insert in the cell
• type (string, optional): The cell type - 'code', 'markdown', or 'raw' (default: 'code')

Returns:

{
  success: true,
  cellType: 'code',
  cellIndex: 5  // 0-based index of the new cell
}

Example usage:

app.commands.execute('notebook:append-execute', {
  source: 'print("Hello from MCP!")',
  type: 'code'
});

notebook:get-selected-cell

Gets information about the currently selected/active cell in the notebook without executing it.

Parameters: None

Returns:

{
  success: true,
  cellType: 'code',           // 'code', 'markdown', or 'raw'
  cellIndex: 3,               // 0-based index
  source: 'print("hello")',   // cell content
  metadata: {...},            // cell metadata
  executionCount: 5           // execution count (code cells only, null otherwise)
}

Example usage:

const cellInfo = await app.commands.execute('notebook:get-selected-cell');
console.log(`Cell ${cellInfo.cellIndex}: ${cellInfo.source}`);

Note: The command is only enabled when a notebook is open and returns an error object if no cell is selected:

{
  success: false,
  error: 'No active cell'
}

Visual Cell Indexing

Installing this extension enhances your JupyterLab experience by displaying cell indices to the left of each cell in orange italics. The goal is to provide a clear and consistent way to reference cells, especially useful when interacting with AI agents through MCP with the jupyter-mcp-server. The tools of the jupyter-mcp-server identify the cells by their index, so having these indices visible helps you when instructing the AI to interact with specific cells.

Requirements

• JupyterLab >= 4.0.0

Install

To install the extension, execute:

pip install jupyter_mcp_tools

Use

make start

Uninstall

To remove the extension, execute:

pip uninstall jupyter_mcp_tools

Troubleshoot

If you are seeing the frontend extension, but it is not working, check that the server extension is enabled:

jupyter server extension list

If the server extension is installed and enabled, but you are not seeing the frontend extension, check the frontend extension is installed:

jupyter labextension list

Contributing

Development install

Note: You will need NodeJS to build the extension package.

The jlpm command is JupyterLab's pinned version of yarn that is installed with JupyterLab. You may use yarn or npm in lieu of jlpm below.

# Clone the repo to your local environment
# Change directory to the jupyter_mcp_tools directory
# Install package in development mode
pip install -e ".[test]"
# Link your development version of the extension with JupyterLab
jupyter labextension develop . --overwrite
# Server extension must be manually installed in develop mode
jupyter server extension enable jupyter_mcp_tools
# Rebuild extension Typescript source after making changes
jlpm build

You can watch the source directory and run JupyterLab at the same time in different terminals to watch for changes in the extension's source and automatically rebuild the extension.

# Watch the source directory in one terminal, automatically rebuilding when needed
jlpm watch
# Run JupyterLab in another terminal
jupyter lab

With the watch command running, every saved change will immediately be built locally and available in your running JupyterLab. Refresh JupyterLab to load the change in your browser (you may need to wait several seconds for the extension to be rebuilt).

By default, the jlpm build command generates the source maps for this extension to make it easier to debug using the browser dev tools. To also generate source maps for the JupyterLab core extensions, you can run the following command:

jupyter lab build --minimize=False

Development uninstall

# Server extension must be manually disabled in develop mode
jupyter server extension disable jupyter_mcp_tools
pip uninstall jupyter_mcp_tools

In development mode, you will also need to remove the symlink created by jupyter labextension develop command. To find its location, you can run jupyter labextension list to figure out where the labextensions folder is located. Then you can remove the symlink named @datalayer/jupyter-mcp-tools within that folder.

Testing the extension

Server tests

This extension is using Pytest for Python code testing.

Install test dependencies (needed only once):

pip install -e ".[test]"
# Each time you install the Python package, you need to restore the front-end extension link
jupyter labextension develop . --overwrite

To execute them, run:

pytest -vv -r ap --cov jupyter_mcp_tools

Frontend tests

This extension is using Jest for JavaScript code testing.

To execute them, execute:

jlpm
jlpm test

Integration tests

This extension uses Playwright for the integration tests (aka user level tests). More precisely, the JupyterLab helper Galata is used to handle testing the extension in JupyterLab.

More information are provided within the ui-tests README.

Packaging the extension

See RELEASE