notebooklm

NotebookLM MCP Server lets you query and manage Google NotebookLM notebooks directly from your AI assistant or MCP-enabled clients. By exposing a set of tools under the notebooklm MCP server, you can list, create, rename, and delete notebooks; attach and manage sources from URLs, local files, or Google Drive; perform grounded notebook queries and initiate research tasks; and generate content through studio-ish capabilities like audio overviews. The server is implemented in TypeScript and follows the MCP (Model Context Protocol), enabling agents to interact with your notebooks as if they were local resources. To use it, start the server (e.g., via npx notebooklm-mcp-server start) and connect your MCP client or editor integration to the notebooklm MCP endpoint. Once connected, you can call tools like notebook_list, notebook_create, notebook_query, notebook_add_url, source_sync, and more to perform tasks against your NotebookLM account.

In practice, you’ll typically authenticate once, then issue commands to manage notebooks and sources, or to run grounded questions against a specific notebook. The tooling split mirrors common notebook workflows: management (notebook_list, notebook_create, notebook_rename, notebook_delete), source management (notebook_add_url, notebook_add_text, notebook_add_local_file, notebook_add_drive, source_delete, source_sync), research and query (notebook_query, research_start, research_poll, research_import), and generation or studio tasks (audio_overview_create, studio_poll). Use the corresponding MCP client or editor extensions to invoke these tools from your environment.

Prerequisites:

• Node.js and npm installed on your system
• Network access to install npm packages

Installation options:

1. Global installation (recommended):

npm install -g notebooklm-mcp-server

Notes: The global install allows you to run the server with a simple start command and enables auto-update behavior at startup.

2. Direct usage with NPX (zero-config):

npx notebooklm-mcp-server auth   # To log in
npx notebooklm-mcp-server start  # To run the server

Authentication:
- Run npx notebooklm-mcp-server auth to authorize your Google account and grant NotebookLM access.
- A browser window will open for login. The session is saved locally for subsequent runs.

Usage:
- Start the server and connect your MCP client/editor to the notebooklm MCP endpoint (default command shown in mcp_config).
- Use tools like notebook_list, notebook_create, notebook_query, and others as described in the Documentation section of the project README.

Tips and common issues:

• First-time authentication requires a browser. Ensure your environment can open external URLs.
• The server stores session data locally. If you encounter authentication errors, try running npx notebooklm-mcp-server refresh_auth to renew the session.
• Auto-update on startup will install newer Google fixes when available. You may be prompted to restart the server after updates.
• If using integrations (Claude Desktop, VS Code, Antigravity, Gemini CLI), follow the exact MCP server command syntax shown in the Quick Start guides to ensure compatibility.
• When working with notebooks and sources, be mindful of destructive actions (notebook_delete, source_delete) and confirm before running.
• Environment variables can be added to customize paths or credentials if the MCP server exposes them in future releases; currently, minimal configuration is typically required.