roomi-fields/notebooklm-mcp

🚀 Two Ways to Use This Server

1️⃣ HTTP REST API (New! Recommended for n8n, Zapier, Make.com)

Use NotebookLM from any tool via HTTP REST API:

Option A: Install from npm

# Install globally
npm install -g @roomi-fields/notebooklm-mcp

# For MCP stdio mode (Claude Code, Cursor)
notebooklm-mcp

# For HTTP server mode
# Note: Currently requires cloning the repo for HTTP mode
# See Option B below

Option B: Install from source (Required for HTTP mode)

# Clone and install
git clone https://github.com/roomi-fields/notebooklm-mcp.git
cd notebooklm-mcp
npm install
npm run build
npm run start:http

Query the API:

curl -X POST http://localhost:3000/ask \
  -H "Content-Type: application/json" \
  -d '{"question": "Explain X", "notebook_id": "my-notebook"}'

Perfect for:

• ✅ n8n workflows and automation
• ✅ Zapier, Make.com integrations
• ✅ Custom web applications
• ✅ Backend APIs

👉 Full HTTP setup guide:

2️⃣ MCP stdio (For Claude Code, Cursor, Codex)

Use NotebookLM directly from your AI coding assistant:

# Claude Code
claude mcp add notebooklm npx @roomi-fields/notebooklm-mcp@latest

# Codex
codex mcp add notebooklm -- npx @roomi-fields/notebooklm-mcp@latest

# Cursor (add to ~/.cursor/mcp.json)
{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@roomi-fields/notebooklm-mcp@latest"]
    }
  }
}

Perfect for:

• ✅ Claude Code, Cursor, Codex
• ✅ Any MCP-compatible AI assistant
• ✅ Direct CLI integration

The Problem

When you tell Claude Code or Cursor to "search through my local documentation", here's what happens:

• Massive token consumption: Searching through documentation means reading multiple files repeatedly
• Inaccurate retrieval: Searches for keywords, misses context and connections between docs
• Hallucinations: When it can't find something, it invents plausible-sounding APIs
• Expensive & slow: Each question requires re-reading multiple files

The Solution

Let your tools chat directly with NotebookLM — Google's zero-hallucination knowledge base powered by Gemini 2.5 that provides intelligent, synthesized answers from your docs.

Your Task → Agent/n8n asks NotebookLM → Gemini synthesizes answer → Correct output

The real advantage: No more manual copy-paste. Your agent/workflow asks NotebookLM directly and gets answers back. Build deep understanding through automatic follow-ups.

Why NotebookLM, Not Local RAG?

What Makes NotebookLM Superior?

1. Pre-processed by Gemini: Upload docs once, get instant expert knowledge
2. Natural language Q&A: Not just retrieval — actual understanding and synthesis
3. Multi-source correlation: Connects information across 50+ documents
4. Citation-backed: Every answer includes source references
5. No infrastructure: No vector DBs, embeddings, or chunking strategies needed

HTTP REST API

Quick Start

# 1. Clone and install
git clone <repo-url> D:\notebooklm-http
cd D:\notebooklm-http
npm install
npm run build

# 2. Configure authentication (one-time)
npm run setup-auth
# Chrome opens → log in with Google → close Chrome

# 3. Start the server

# Option A: Foreground mode (terminal stays open)
npm run start:http

# Option B: Background daemon mode (recommended for production)
npm run daemon:start    # Start in background
npm run daemon:logs     # View logs
npm run daemon:status   # Check status
npm run daemon:stop     # Stop server

# 4. Add your first notebook (in another terminal if using Option A)
curl -X POST http://localhost:3000/notebooks \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://notebooklm.google.com/notebook/YOUR-NOTEBOOK-ID",
    "name": "My Knowledge Base",
    "description": "My documentation",
    "topics": ["docs", "api"]
  }'

# 5. Query NotebookLM
curl -X POST http://localhost:3000/ask \
  -H "Content-Type: application/json" \
  -d '{"question": "Explain X"}'

# 6. Validate installation (optional)
.\deployment\scripts\test-server.ps1

👉 See for automated test suite

API Endpoints

👉 Full API documentation:

n8n Integration

Perfect for n8n workflows:

{
  "nodes": [
    {
      "name": "Ask NotebookLM",
      "type": "n8n-nodes-base.httpRequest",
      "parameters": {
        "method": "POST",
        "url": "http://your-server:3000/ask",
        "jsonParameters": true,
        "bodyParametersJson": {
          "question": "{{ $json.query }}",
          "notebook_id": "my-notebook"
        }
      }
    }
  ]
}

👉 n8n guide:

Background Daemon Mode

Run the HTTP server as a background process without keeping a terminal window open:

# Start server in background
npm run daemon:start

# View real-time logs
npm run daemon:logs

# Check server status
npm run daemon:status

# Restart server
npm run daemon:restart

# Stop server
npm run daemon:stop

# Remove from PM2 process list
npm run daemon:delete

Features:

• ✅ Runs in background without terminal window
• ✅ Auto-restart on crash
• ✅ Centralized log management (logs/pm2-*.log)
• ✅ Memory limit protection (1GB max)
• ✅ Production-ready process management

Configuration: Edit ecosystem.config.cjs to customize PM2 behavior (env vars, restart policy, etc.)

Installation & Documentation

• 📖 — Step-by-step setup
• 🔧 — Environment variables, security
• 📡 — Complete endpoint documentation
• 📚 — Multi-notebook management
• ✅ — Automated validation scripts
• 🐛 — Common issues

MCP Installation

claude mcp add notebooklm npx @roomi-fields/notebooklm-mcp@latest

codex mcp add notebooklm -- npx @roomi-fields/notebooklm-mcp@latest

gemini mcp add notebooklm npx @roomi-fields/notebooklm-mcp@latest

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["-y", "@roomi-fields/notebooklm-mcp@latest"]
    }
  }
}

amp mcp add notebooklm -- npx @roomi-fields/notebooklm-mcp@latest

code --add-mcp '{"name":"notebooklm","command":"npx","args":["@roomi-fields/notebooklm-mcp@latest"]}'

{
  "mcpServers": {
    "notebooklm": {
      "command": "npx",
      "args": ["@roomi-fields/notebooklm-mcp@latest"]
    }
  }
}

MCP Quick Start

1. Install the MCP server (see above)

2. Authenticate (one-time)

Say in your chat (Claude/Codex):

"Log me in to NotebookLM"

A Chrome window opens → log in with Google

3. Create your knowledge base

Go to notebooklm.google.com → Create notebook → Upload your docs:

• 📄 PDFs, Google Docs, markdown files
• 🔗 Websites, GitHub repos
• 🎥 YouTube videos
• 📚 Multiple sources per notebook

Share: ⚙️ Share → Anyone with link → Copy

4. Let Claude use it

"I'm building with [library]. Here's my NotebookLM: [link]"

That's it. Claude now asks NotebookLM whatever it needs, building expertise before writing code.

Related Project: Claude Code Skill (by original author)

The original author PleasePrompto also created a Python-based Claude Code Skill as an alternative approach:

🔗 NotebookLM Claude Code Skill - Python skill for Claude Code

When to use which approach?

Which one should you choose?

• Use this MCP project if you want:

  • Persistent sessions (faster repeated queries)
  • Compatibility with multiple tools (Cursor, Codex, etc.)
  • HTTP REST API for n8n/Zapier automation
  • TypeScript-based development

• Use the original Skill if you prefer:

  • Python-based workflow
  • Simpler clone-and-use installation
  • Stateless queries (no session management)
  • Only using Claude Code locally

Both use the same Patchright browser automation technology and provide zero-hallucination answers from NotebookLM.

Real-World Example

Building an n8n Workflow Without Hallucinations

Challenge: n8n's API is new — Claude hallucinates node names and functions.

Solution:

1. Downloaded complete n8n documentation → merged into manageable chunks
2. Uploaded to NotebookLM
3. Told Claude: "Build me a Gmail spam filter workflow. Use this NotebookLM: [link]"

Watch the AI-to-AI conversation:

Claude → "How does Gmail integration work in n8n?"
NotebookLM → "Use Gmail Trigger with polling, or Gmail node with Get Many..."

Claude → "How to decode base64 email body?"
NotebookLM → "Body is base64url encoded in payload.parts, use Function node..."

Claude → "How to parse OpenAI response as JSON?"
NotebookLM → "Set responseFormat to json, use {{ $json.spam }} in IF node..."

Claude → "What about error handling if the API fails?"
NotebookLM → "Use Error Trigger node with Continue On Fail enabled..."

Claude → ✅ "Here's your complete workflow JSON..."

Result: Perfect workflow on first try. No debugging hallucinated APIs.

Core Features

Zero Hallucinations

NotebookLM refuses to answer if information isn't in your docs. No invented APIs.

Multi-Notebook Library

Manage multiple NotebookLM notebooks with automatic validation, duplicate detection, and smart selection.

Autonomous Research

Claude asks follow-up questions automatically, building complete understanding before coding.

Deep, Iterative Research

• Claude automatically asks follow-up questions to build complete understanding
• Each answer triggers deeper questions until Claude has all the details
• Example: For n8n workflow, Claude asked multiple sequential questions about Gmail integration, error handling, and data transformation

HTTP REST API

Use NotebookLM from n8n, Zapier, Make.com, or any HTTP client. No MCP required.

Cross-Tool Sharing

Set up once, use everywhere. Claude Code, Codex, Cursor, n8n — all can access the same library.

Architecture

graph LR
    A[Your Task] --> B[Claude/n8n/HTTP Client]
    B --> C[MCP/HTTP Server]
    C --> D[Chrome Automation]
    D --> E[NotebookLM]
    E --> F[Gemini 2.5]
    F --> G[Your Docs]
    G --> F
    F --> E
    E --> D
    D --> C
    C --> B
    B --> H[Accurate Output]

Common Commands (MCP Mode)

Comparison to Alternatives

vs. Downloading docs locally

• You: Download docs → Claude: "search through these files"
• Problem: Claude reads thousands of files → massive token usage, often misses connections
• NotebookLM: Pre-indexed by Gemini, semantic understanding across all docs

vs. Web search

• You: "Research X online"
• Problem: Outdated info, hallucinated examples, unreliable sources
• NotebookLM: Only your trusted docs, always current, with citations

vs. Local RAG setup

• You: Set up embeddings, vector DB, chunking strategy, retrieval pipeline
• Problem: Hours of setup, tuning retrieval, still gets "creative" with gaps
• NotebookLM: Upload docs → done. Google handles everything.

FAQ

Is it really zero hallucinations? Yes. NotebookLM is specifically designed to only answer from uploaded sources. If it doesn't know, it says so.

What about rate limits? Free tier has daily query limits per Google account. Quick account switching supported for continued research.

How secure is this? Chrome runs locally. Your credentials never leave your machine. Use a dedicated Google account if concerned.

Can I see what's happening? Yes! Say "Show me the browser" (MCP mode) or set HEADLESS=false (HTTP mode) to watch the live NotebookLM conversation.

What makes this better than Claude's built-in knowledge? Your docs are always current. No training cutoff. No hallucinations. Perfect for new libraries, internal APIs, or fast-moving projects.

The Bottom Line

Without NotebookLM: Write code → Find it's wrong → Debug hallucinated APIs → Repeat

With NotebookLM: Research first → Write correct code → Ship faster

Stop debugging hallucinations. Start shipping accurate code.

Disclaimer

This tool automates browser interactions with NotebookLM to make your workflow more efficient. However, a few friendly reminders:

About browser automation: While I've built in humanization features (realistic typing speeds, natural delays, mouse movements) to make the automation behave more naturally, I can't guarantee Google won't detect or flag automated usage. I recommend using a dedicated Google account for automation rather than your primary account—think of it like web scraping: probably fine, but better safe than sorry!

About CLI tools and AI agents: CLI tools like Claude Code, Codex, and similar AI-powered assistants are incredibly powerful, but they can make mistakes. Please use them with care and awareness:

• Always review changes before committing or deploying
• Test in safe environments first
• Keep backups of important work
• Remember: AI agents are assistants, not infallible oracles

I built this tool for myself because I was tired of the copy-paste dance between NotebookLM and my editor. I'm sharing it in the hope it helps others too, but I can't take responsibility for any issues, data loss, or account problems that might occur. Use at your own discretion and judgment.

That said, if you run into problems or have questions, feel free to open an issue on GitHub. I'm happy to help troubleshoot!

Roadmap

Future features planned for upcoming releases:

🚀 Stateless Server Mode

Goal: Run the HTTP server without keeping a terminal window open permanently.

Planned options:

• PM2 integration (Recommended): Cross-platform process manager with auto-restart, monitoring, and logs
  • Simple setup: npm run daemon:start (uses PM2 with optimized config)
  • Automatic startup on system boot
  • Built-in log rotation and monitoring
• Windows Service: Native Windows service installation via nssm or node-windows
• Serverless-ready mode: Lambda/cold-start compatible with lazy browser session initialization

Status: Planned for v1.2.0

🤖 Auto-fill Notebook Metadata

Goal: Automatically generate notebook name, description, and keywords when adding a notebook.

How it works:

1. When adding a notebook with empty metadata, automatically ask NotebookLM:
   • "If you had to name this notebook in one word, what would it be?"
   • "Give me 10 relevant keywords for this content"
   • "Describe this notebook in one sentence"
2. Parse the response and auto-populate metadata fields
3. Available as both automatic mode (on add) and manual command (/notebooks/:id/auto-fill)

Benefits:

• No more manual metadata entry
• Consistent, AI-generated descriptions
• Better notebook organization and searchability

Status: Planned for v1.2.0 or v1.3.0

💡 Have an idea?

Open a discussion to suggest new features!

Contributing

Found a bug? Have a feature idea? Open an issue or submit a PR!

See for contribution guidelines.

License

MIT — Use freely in your projects.

See for details.