Skip to content

NotebookLM MCP Setup Instructions

Status

Requires interactive browser authentication - cannot be automated.

Current State

  • MCP tool available: mcp__notebooklm__setup_auth
  • Config directory exists: ~/.config/notebooklm-mcp/ (empty)
  • 1 notebook configured: "Claude Developer Documentation"
  • Auth status: Not authenticated

Setup Steps (Manual)

  1. Run setup in a Claude session with browser access: 

    Call: mcp__notebooklm__setup_auth(show_browser=true)

  2. Complete Google login in the browser window that opens

  3. Verify authentication: 

    Call: mcp__notebooklm__get_health

  4. Test notebook access: 

    Call: mcp__notebooklm__list_notebooks

Available Tools (After Auth)

Notebook Management

  • mcp__notebooklm__create_notebook - Create new notebook
  • mcp__notebooklm__add_notebook - Add existing notebook to library
  • mcp__notebooklm__remove_notebook - Remove from library
  • mcp__notebooklm__get_notebook - Get notebook details
  • mcp__notebooklm__list_notebooks - List all notebooks

Sources

  • mcp__notebooklm__add_source - Add URL/text/YouTube source
  • mcp__notebooklm__add_sources_bulk - Add multiple sources
  • mcp__notebooklm__list_notebook_sources - List sources in notebook

Querying

  • mcp__notebooklm__ask_question - RAG query on notebook sources
  • mcp__notebooklm__create_note - Create note in notebook

Management

  • mcp__notebooklm__rename_notebook - Rename notebook
  • mcp__notebooklm__rename_source - Rename source
  • mcp__notebooklm__delete_source - Delete source

Audio

  • mcp__notebooklm__generate_audio_overview - Generate podcast
  • mcp__notebooklm__get_audio_overview_status - Check audio status
  • mcp__notebooklm__download_audio_overview - Download audio file

Import

  • mcp__notebooklm__import_youtube_playlist - Import YouTube playlist
  • mcp__notebooklm__import_rss_feed - Import RSS feed

Use Cases for Aegis

1. Research Knowledge Base

Create notebooks for: - Competitive intelligence (Reddit, HN, ProductHunt) - Technical documentation (Claude API, frameworks) - Market research (SaaS pricing, trends)

2. Meeting Notes

Automatically transcribe and synthesize: - Discord meeting summaries - Project planning sessions - Code review discussions

RAG search across: - Code repositories - Project documentation - Architecture decisions

Integration Opportunities

With Aegis Memory

NotebookLM.ask_question() → Knowledge Graph
- Extract entities from notebook answers
- Add to Graphiti for cross-referencing

With Proactive Scanner

Scanner → NotebookLM.add_source()
- Add relevant articles to research notebooks
- Auto-generate briefings from RAG queries

With Discord

Discord messages → NotebookLM.add_source()
- Build knowledge base from community discussions
- Query for trends and patterns

Notebooks to Create

  1. Competitive Intelligence
  2. Sources: Reddit posts, HN discussions, Product Hunt launches

  3. Query: "What are competitors charging for API access?"

  4. Technical Documentation

  5. Sources: Claude API docs, framework guides

  6. Query: "How do I implement streaming responses?"

  7. Market Research

  8. Sources: Industry reports, pricing studies
  9. Query: "What are common SaaS pricing models?"

Estimated Setup Time

  • 5 minutes for initial auth
  • 10 minutes per notebook to add sources
  • 1-2 hours to build comprehensive knowledge base

Next Steps (When Auth is Complete)

  1. Create research notebook for Intel Dashboard competitors
  2. Add sources from Reddit demand signal research
  3. Test RAG queries for pricing insights
  4. Integrate with daily digest workflow