NotebookLM MCP Server

This repository provides a Model Context Protocol (MCP) server for NotebookLM. It allows AI assistants (like Claude, Antigravity, or others supporting MCP) to interact with your NotebookLM notebooks, sources, and conversations.

Features

• List Notebooks: View all your NotebookLM notebooks.
• Create Notebooks: Programmatically create new notebooks.
• Manage Sources: Add websites, Google Drive documents, or pasted text to your notebooks.
• Query Notebooks: Ask questions about your sources using the NotebookLM AI.
• Conversation History: Full support for follow-up questions and conversation context.
• Auto-Save Notes: Automatically save AI responses as notes in your notebooks.

Prerequisites

• Python 3.10 or higher.
• A Google Account with access to NotebookLM.
• An MCP-compatible client (e.g., Cursor, Claude Desktop).

Quick Start

1. Clone the repository

git clone https://github.com/YOUR_USERNAME/MCPNotebookLM.git
cd MCPNotebookLM

2. Set up the environment

Install dependencies:

pip install -r requirements.txt

Authenticate with NotebookLM:

notebooklm-mcp-auth

Follow the prompts to authorize the application. This will create a local auth.json file in ~/.notebooklm-mcp/.

3. Configure your MCP Client

For Cursor:

1. Copy mcp_config.json.example to your Cursor config directory:

   • Linux: ~/.config/cursor/mcp.json
   • macOS: ~/Library/Application Support/Cursor/mcp.json
   • Windows: %APPDATA%\Cursor\mcp.json

2. Edit the command field with the absolute path to notebooklm-mcp:

   {
     "mcpServers": {
       "notebooklm": {
         "command": "/home/YOUR_USER/.local/bin/notebooklm-mcp",
         "args": [],
         "env": {}
       }
     }
   }
   

3. Find the binary path:

   which notebooklm-mcp
   # or
   ls ~/.local/bin/notebooklm-mcp
   

4. Restart Cursor to apply the configuration.

Usage Examples

Basic Usage

Test your setup by listing notebooks:

python3 query_notebook_mcp.py

Or query a notebook directly:

python3 query_notebook_mcp.py <notebook_id> "Your question"

Auto-Save Notes Feature

The repository includes an automatic note-saving feature that saves all AI responses as notes in your notebooks. This is especially useful when working through MCP API, as responses aren't automatically saved in the web interface history.

Quick start:

from auto_save_notes import query_and_save

answer, source_id = query_and_save(
    notebook_id="your-notebook-id",
    question="What is Python?",
    auto_save=True
)

See docs/AUTO_SAVE_NOTES.md for detailed documentation (if available locally).

Security Note

[!WARNING] Your authentication tokens are stored locally in ~/.notebooklm-mcp/auth.json. Never share this file or commit it to a public repository. The .gitignore in this repo is already configured to ignore this folder.

License

MIT