NotebookLM
STDIOGoogle NotebookLM自动化MCP服务器
Google NotebookLM自动化MCP服务器
Professional MCP server for Google NotebookLM automation • Available on PyPI • Production Ready
# Install UV (modern Python package manager) curl -LsSf https://astral.sh/uv/install.sh | sh # Install NotebookLM MCP from PyPI uv add notebooklm-mcp # Initialize with your NotebookLM URL uv run notebooklm-mcp init https://notebooklm.google.com/notebook/YOUR_NOTEBOOK_ID
What happens after init:
notebooklm-config.json with your settingschrome_profile_notebooklm/ folder for persistent authentication# Start server (STDIO for MCP clients) uv run notebooklm-mcp --config notebooklm-config.json server # Start HTTP server for web testing uv run notebooklm-mcp --config notebooklm-config.json server --transport http --port 8001 # Interactive chat mode uv run notebooklm-mcp --config notebooklm-config.json chat --message "Who are you ?"
If you're contributing to this project, check out our Taskfile for enhanced developer experience:
git clone https://github.com/khengyun/notebooklm-mcp.git cd notebooklm-mcp # Complete setup with development tools task setup # Show all available development tasks task --list
If you prefer pip over UV:
# Install with pip pip install notebooklm-mcp # Initialize notebooklm-mcp init https://notebooklm.google.com/notebook/YOUR_NOTEBOOK_ID # Start server notebooklm-mcp --config notebooklm-config.json server
After running init, your working directory will contain:
your-project/ ├── notebooklm-config.json # Configuration file ├── chrome_profile_notebooklm/ # Browser profile (persistent auth) │ ├── Default/ # Chrome profile data │ ├── SingletonSocket # Session files │ └── ... # Other Chrome data └── your-other-files
Key files:
notebooklm-config.json: Contains notebook ID, server settings, auth configurationchrome_profile_notebooklm/: Stores Google authentication session (enables headless operation)| Tool | Description | Parameters | 
|---|---|---|
healthcheck | Server health status | None | 
send_chat_message | Send message to NotebookLM | message: str, wait_for_response: bool | 
get_chat_response | Get response with timeout | timeout: int | 
chat_with_notebook | Complete interaction | message: str, notebook_id?: str | 
navigate_to_notebook | Switch notebooks | notebook_id: str | 
get_default_notebook | Current notebook | None | 
set_default_notebook | Set default | notebook_id: str | 
get_quick_response | Instant response | None | 
For contributors and advanced users who want enhanced productivity, we provide a comprehensive Taskfile with 20+ automation tasks:
# 📦 Dependency Management task deps-add -- requests # Add dependency task deps-add-dev -- pytest # Add dev dependency task deps-remove -- requests # Remove dependency task deps-list # List dependencies task deps-update # Update all dependencies # 🧪 Testing & Quality task test # Run all tests task test-quick # Quick validation test task test-coverage # Coverage analysis task enforce-test # MANDATORY after function changes task lint # Run all linting task format # Format code (Black + isort + Ruff) # 🏗️ Build & Release task build # Build package task clean # Clean artifacts # 🚀 Server Commands task server-stdio # STDIO server task server-http # HTTP server task server-sse # SSE server # Show all available tasks task --list
💡 Pro Tip: Install Task for the best developer experience:
go install github.com/go-task/task/v3/cmd/task@latest
task server-stdio # For: LangGraph, CrewAI, AutoGen
task server-http # Access: http://localhost:8001/mcp # For: Web testing, REST APIs
task server-sse # Access: http://localhost:8002/ # For: Real-time streaming
from fastmcp import Client from fastmcp.client.transports import StreamableHttpTransport transport = StreamableHttpTransport(url="http://localhost:8001/mcp") async with Client(transport) as client: tools = await client.list_tools() result = await client.call_tool("healthcheck", {})
# Test with curl curl -X POST http://localhost:8001/mcp \ -H "Content-Type: application/json" \ -H "Accept: application/json, text/event-stream" \ -d '{"jsonrpc": "2.0", "id": 1, "method": "tools/list", "params": {}}'
from fastmcp import Client from fastmcp.client.transports import StreamableHttpTransport # HTTP transport transport = StreamableHttpTransport(url="http://localhost:8001/mcp") client = Client(transport) tools = await client.list_tools()
from crewai_tools import BaseTool from fastmcp import Client class NotebookLMTool(BaseTool): name = "notebooklm" description = "Chat with NotebookLM" async def _arun(self, message: str): client = Client("http://localhost:8001/mcp") result = await client.call_tool("chat_with_notebook", {"message": message}) return result
# First time - opens browser for login notebooklm-mcp init https://notebooklm.google.com/notebook/abc123 # Subsequent runs - uses saved session notebooklm-mcp --config notebooklm-config.json server
# Interactive browser login notebooklm-mcp --config notebooklm-config.json server # Check connection notebooklm-mcp --config notebooklm-config.json test --notebook YOUR_NOTEBOOK_ID
# 1. Generate config + Chrome profile with guided login uv run notebooklm-mcp init https://notebooklm.google.com/notebook/YOUR_NOTEBOOK_ID # If you installed via pip, run: notebooklm-mcp init https://notebooklm.google.com/notebook/YOUR_NOTEBOOK_ID # 2. Build the container image docker build -t notebooklm-mcp . # 3. Run the server with your mounted config/profile docker run -d \ --name notebooklm-mcp \ --restart unless-stopped \ -v $(pwd)/notebooklm-config.json:/app/notebooklm-config.json:ro \ -v $(pwd)/chrome_profile_notebooklm:/app/chrome_profile_notebooklm \ notebooklm-mcp:latest
version: '3.8' services: notebooklm-mcp: image: notebooklm-mcp:latest build: . restart: unless-stopped volumes: - ./notebooklm-config.json:/app/notebooklm-config.json:ro - ./chrome_profile_notebooklm:/app/chrome_profile_notebooklm
Start the stack with docker compose up -d after running the init command once so
that notebooklm-config.json and chrome_profile_notebooklm/ exist. The server
runs in STDIO mode by default; uncomment the HTTP/SSE ports in
docker-compose.yml if your client requires them.
notebooklm-config.json){ "default_notebook_id": "your-notebook-id", "headless": true, "timeout": 30, "auth": { "profile_dir": "./chrome_profile_notebooklm" }, "debug": false }
export NOTEBOOKLM_NOTEBOOK_ID="your-notebook-id" export NOTEBOOKLM_HEADLESS=true export NOTEBOOKLM_DEBUG=false
| Feature | Traditional MCP | FastMCP v2 | 
|---|---|---|
| Tool registration | Manual schema | Auto-generated | 
| Type validation | Manual | Automatic | 
| Error handling | Basic | Enhanced | 
| Development speed | Standard | 5x faster | 
| HTTP support | Limited | Full | 
git clone https://github.com/khengyun/notebooklm-mcp cd notebooklm-mcp # With UV (recommended) uv sync --all-groups # Or with pip pip install -e ".[dev]"
# Run tests with UV uv run pytest # With coverage uv run pytest --cov=notebooklm_mcp # Integration tests uv run pytest tests/test_integration.py # Or use Taskfile for development task test task test-coverage
# Format code with UV uv run black src/ tests/ uv run ruff check src/ tests/ # Type checking uv run mypy src/ # Or use Taskfile shortcuts task format task lint
MIT License - see LICENSE file for details.
Built with ❤️ using FastMCP v2 - Modern MCP development made simple!