This project provides a Middleware Control Plane (MCP) server for Google's NotebookLM, enabling programmatic interaction with its AI-powered research and summarization features. It targets developers and power users seeking to integrate NotebookLM's capabilities into custom workflows and AI assistant tools, offering automation for tasks like notebook management, content ingestion, AI analysis, and content generation.

How It Works

The MCP server exposes a suite of 31 tools that interface with NotebookLM's internal, undocumented APIs. It relies on authentication via browser cookies, which can be automatically extracted using a provided tool that launches a dedicated Chrome instance, or manually extracted. This architecture allows various AI tools and CLIs (e.g., Claude Code, Cursor, Gemini CLI) to control and leverage NotebookLM functionalities through a unified interface.

Quick Start & Requirements

• Installation: Install via PyPI (uv tool install notebooklm-mcp-server, pip install notebooklm-mcp-server, pipx install notebooklm-mcp-server) or from source.
• Prerequisites: Python, a web browser (Chrome recommended for auto-authentication). pnpm may be needed for specific troubleshooting.
• Authentication: Run notebooklm-mcp-auth for automatic cookie extraction via Chrome, or notebooklm-mcp-auth --file for manual extraction.
• Configuration: Requires adding the MCP server configuration to specific JSON files for supported AI tools (e.g., ~/.claude.json, ~/.cursor/mcp.json).
• Resource Footprint: The MCP's 31 tools consume significant context window space; disabling it when not in use is recommended.

Highlighted Details

• Full lifecycle management for NotebookLM notebooks: create, list, get details, rename, delete.
• Versatile source ingestion: add content via URL, YouTube links, pasted text, or Google Drive documents.
• Advanced AI features: generate summaries, keywords, AI-driven research, and content analysis across sources.
• Content generation capabilities: create audio podcasts, video overviews, slide decks, infographics, and mind maps from notebook content.
• Automated research initiation and source discovery/import from web or Google Drive.
• Built-in mechanisms for refreshing authentication tokens and syncing stale Drive sources.

Maintenance & Community

The project explicitly acknowledges its origin as an AI-assisted development effort and actively encourages contributions from experienced developers for refactoring, optimization, and architectural improvements. No specific community channels (e.g., Discord, Slack) are listed.

Licensing & Compatibility

The project is released under the MIT License, permitting broad use, modification, and distribution, including for commercial purposes, with minimal restrictions.

Limitations & Caveats

The MCP interacts with undocumented internal APIs that are subject to change without notice. The free tier of NotebookLM imposes rate limits (approximately 50 queries per day). Authentication relies on browser cookies that expire roughly every 2-4 weeks, requiring periodic re-authentication. The codebase's development methodology may result in non-standard patterns or optimizations.