NotebookLM MCP Server
Enables AI agents to programmatically access Google NotebookLM through browser automation for managing notebooks, sources, and chat interactions. It supports automated content generation including audio overviews, study guides, and quizzes directly within AI workflows.
README
NotebookLM MCP Server
A remote Model Context Protocol (MCP) server that gives AI agents (Claude, Cursor, Copilot, etc.) full programmatic access to Google NotebookLM via browser automation.
Hosted on Vercel Β· Browser powered by Browserless.io
πΊοΈ Architecture
AI Agent (Claude / Cursor)
β MCP (Streamable HTTP)
βΌ
βββββββββββββββββββββββ
β Vercel Function β β Next.js App Router
β (TypeScript / MCP) β mcp-handler adapter
βββββββββββ¬ββββββββββββ
β WebSocket (CDP)
βΌ
βββββββββββββββββββββββ
β Browserless.io β β Remote headless Chromium
β (free tier) β stealth mode enabled
βββββββββββ¬ββββββββββββ
β Playwright
βΌ
βββββββββββββββββββββββ
β Google NotebookLM β β Authenticated via stored cookies
βββββββββββββββββββββββ
β‘ Quick Start
Prerequisites
- Node.js 18+
- A free Browserless.io account β get your API token
- A Google account with access to NotebookLM
- A Vercel account
Step 1 β Clone & install
git clone <your-repo-url>
cd notebooklm-mcp
npm install
Step 2 β Extract Google cookies (one-time)
This step opens a browser window for you to sign in. Runs locally only.
# Install playwright for local cookie extraction
npm install playwright --save-dev
npx playwright install chromium
# Run the extractor
npm run get-cookies
- A Chrome browser will open β sign in to your Google account
- Navigate to NotebookLM and verify you can see your notebooks
- Switch back to the terminal and press Enter
- The script prints a
NOTEBOOKLM_STORAGE_STATE=...value β copy it
Step 3 β Configure environment variables
Copy .env.example to .env.local and fill in your values:
cp .env.example .env.local
# From https://browserless.io β your dashboard
BROWSERLESS_TOKEN=your_token_here
# From the get-cookies script output
NOTEBOOKLM_STORAGE_STATE=eyJjb29...
# Optional: closest region (sfo | lon | sea | fra)
BROWSERLESS_REGION=sfo
Step 4 β Test locally
npm run dev
# MCP endpoint: http://localhost:3000/api/mcp
Step 5 β Deploy to Vercel
# Install Vercel CLI
npm i -g vercel
# Deploy (follow the prompts)
vercel
# Add your environment variables (or use the Vercel dashboard)
vercel env add BROWSERLESS_TOKEN
vercel env add NOTEBOOKLM_STORAGE_STATE
vercel env add BROWSERLESS_REGION
# Deploy to production
vercel --prod
Your MCP server will be live at:
https://your-project.vercel.app/api/mcp
Step 6 β Connect your AI agent
Claude Desktop (~/Library/Application\ Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"notebooklm": {
"url": "https://your-project.vercel.app/api/mcp"
}
}
}
Cursor (.cursor/mcp.json):
{
"mcpServers": {
"notebooklm": {
"url": "https://your-project.vercel.app/api/mcp"
}
}
}
π οΈ Available MCP Tools
Notebooks
| Tool | Description |
|---|---|
notebooklm_list_notebooks |
List all notebooks in your account |
notebooklm_create_notebook |
Create a new notebook |
notebooklm_get_notebook |
Get details of a notebook by ID |
notebooklm_delete_notebook |
Delete a notebook permanently |
Sources
| Tool | Description |
|---|---|
notebooklm_list_sources |
List all sources in a notebook |
notebooklm_add_source_url |
Add a website URL as a source |
notebooklm_add_source_youtube |
Add a YouTube video as a source |
notebooklm_add_source_text |
Add plain text as a source |
notebooklm_remove_source |
Remove a source from a notebook |
Chat / Query
| Tool | Description |
|---|---|
notebooklm_ask |
Ask a question (single turn) |
notebooklm_conversation |
Multi-turn conversation (up to 10 questions) |
Content Generation
| Tool | Description |
|---|---|
notebooklm_generate_audio |
Generate an Audio Overview (podcast) |
notebooklm_generate_quiz |
Generate a quiz |
notebooklm_generate_flashcards |
Generate flashcards |
notebooklm_generate_mindmap |
Generate a mind map |
notebooklm_generate_slideshow |
Generate a slide deck |
notebooklm_generate_study_guide |
Generate a study guide |
notebooklm_generate_briefing |
Generate a briefing document |
Utility
| Tool | Description |
|---|---|
notebooklm_health_check |
Verify Browserless.io connection and session validity |
π Session Management
Google sessions typically last 30β90 days. When your session expires:
- Run
npm run get-cookiesagain on your local machine - Copy the new
NOTEBOOKLM_STORAGE_STATEvalue - Update it in Vercel:
vercel env rm NOTEBOOKLM_STORAGE_STATE && vercel env add NOTEBOOKLM_STORAGE_STATE - Re-deploy:
vercel --prod
The notebooklm_health_check tool will tell you if the session is still valid.
π° Browserless.io Free Tier
The free tier gives you ~1,000 units/month.
| Operation | Approx. cost |
|---|---|
| List notebooks | ~1 unit |
| Add a URL source | ~2β3 units |
| Ask a question | ~2 units |
| Generate audio | ~2 units |
For typical personal use (~50β100 tool calls/month), the free tier is more than sufficient.
β οΈ Limitations
- Vercel Hobby timeout: Functions are limited to 60 seconds. Long-running operations (audio generation) are started asynchronously and must be checked in the NotebookLM UI.
- Session expiry: Google cookies expire. Re-run
get-cookieswhen authentication fails. - UI changes: NotebookLM is a product under active development. If selectors break, open an issue.
- Unofficial: This uses browser automation, not an official API. Google could change NotebookLM's interface at any time.
π§ Troubleshooting
"Session has expired" error
β Re-run npm run get-cookies and update NOTEBOOKLM_STORAGE_STATE in Vercel.
"BROWSERLESS_TOKEN is not set" error β Add the token in Vercel dashboard β Settings β Environment Variables.
Tool returns "Could not find button" error β NotebookLM's UI may have been updated. Check the notebooklm-py repo for updated selectors.
Timeout errors β Browserless.io free tier may be under load. Wait a few minutes and retry.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.