NotebookLM MCP
Gives Claude Desktop full control over Google NotebookLM via 56 MCP tools — manage notebooks, sources, chat, artifacts, notes, sharing, research, and export.
README
NotebookLM MCP for Claude Desktop
Gives Claude Desktop full control over Google NotebookLM via 56 MCP tools — manage notebooks, sources, chat, artifacts, notes, sharing, research, and export.
| Category | Example tools |
|---|---|
| Notebooks | list, create, get, rename, delete, summary, description |
| Sources | list, add (URL/text/file/Drive), get, fulltext, guide, delete, refresh, rename |
| Chat | ask (with citations), configure, set mode, history |
| Artifacts | generate audio, video, slides, quiz, flashcards, report, infographic, mind map, study guide, data table |
| Artifact Mgmt | list, poll status, wait, download, delete, rename, revise slide, suggest reports |
| Notes | create, list, update, delete |
| Sharing | get status, set public/private, add/remove users, set view level, get share URL |
| Research | start web/Drive research, poll status, import sources |
| Settings | get/set output language |
| Export | export to Google Docs/Sheets |
Prerequisites
- Python 3.10 or higher
- A Google account with NotebookLM access (notebooklm.google.com)
- Git (optional — you can also download a ZIP)
Setup
1. Get the code
git clone https://github.com/sprmdta/Experiment_NotebookM_MCP.git
cd Experiment_NotebookM_MCP
Or download and extract the ZIP, then open a terminal in the extracted folder.
2. Create a virtual environment
python -m venv .venv
3. Install dependencies
Windows:
.venv\Scripts\pip install -r requirements.txt
Mac/Linux:
.venv/bin/pip install -r requirements.txt
4. Install the Playwright browser
Windows:
.venv\Scripts\playwright install chromium
Mac/Linux:
.venv/bin/playwright install chromium
5. Log in to Google
Run the login command from inside the project folder:
Windows (PowerShell or Command Prompt):
cd "C:\YOUR_PATH_TO_REPO\Experiment_NotebookM_MCP"
.venv\Scripts\python -m notebooklm login
Mac/Linux:
cd /YOUR_PATH_TO_REPO/Experiment_NotebookM_MCP
.venv/bin/python -m notebooklm login
What happens during login
The login command is interactive — it opens a browser and waits for you to confirm before saving anything. Follow these steps exactly:
- A Chromium browser window opens automatically and navigates to NotebookLM
- Sign in with the Google account linked to your NotebookLM account
- Complete any 2-factor authentication if prompted
- Wait until you can see the NotebookLM homepage (your list of notebooks)
- Switch back to your terminal window
- Press ENTER in the terminal to save the session
Critical: Do NOT close the browser window before pressing ENTER. Closing the browser exits the process without saving. The session cookies are only written to disk when you press ENTER in the terminal.
What success looks like
After pressing ENTER, the terminal will print:
Authentication saved to: C:\Users\YOUR_USERNAME\.notebooklm\storage_state.json
This file (storage_state.json) contains your Google session cookies. The MCP server reads this file every time it makes a request to NotebookLM.
Verify the login worked
You can confirm the file was created:
Windows:
dir "%USERPROFILE%\.notebooklm\storage_state.json"
Mac/Linux:
ls ~/.notebooklm/storage_state.json
If the file exists, login succeeded. If it does not exist, the login was not completed — re-run the command and make sure to press ENTER in the terminal.
6. Configure Claude Desktop
Open (or create) the Claude Desktop config file:
- Windows:
C:\Users\YOUR_USERNAME\AppData\Roaming\Claude\claude_desktop_config.json - Mac:
~/Library/Application Support/Claude/claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
Windows (replace YOUR_PATH_TO_REPO with the actual path where you cloned the repo):
{
"mcpServers": {
"notebooklm": {
"command": "C:\\YOUR_PATH_TO_REPO\\Experiment_NotebookM_MCP\\.venv\\Scripts\\python.exe",
"args": ["C:\\YOUR_PATH_TO_REPO\\Experiment_NotebookM_MCP\\notebooklm_mcp.py"],
"env": {
"NOTEBOOKLM_STORAGE_PATH": "C:\\Users\\YOUR_USERNAME\\.notebooklm"
}
}
}
}
Mac/Linux:
{
"mcpServers": {
"notebooklm": {
"command": "/YOUR_PATH_TO_REPO/Experiment_NotebookM_MCP/.venv/bin/python",
"args": ["/YOUR_PATH_TO_REPO/Experiment_NotebookM_MCP/notebooklm_mcp.py"],
"env": {
"NOTEBOOKLM_STORAGE_PATH": "/Users/YOUR_USERNAME/.notebooklm"
}
}
}
}
7. Restart Claude Desktop
Fully quit Claude Desktop — closing the window is not enough. Right-click the Claude icon in the system tray (bottom-right taskbar on Windows, menu bar on Mac) and choose Quit, then reopen it.
Claude Desktop reads the config only at startup. If you edit the config while it is running, the changes will not take effect until you fully restart it.
The NotebookLM tools appear in the tools panel (hammer icon) once connected.
8. Test
Ask Claude: List my NotebookLM notebooks
Claude will call notebooklm_list_notebooks and return your notebooks.
Re-authentication
Session cookies expire periodically (typically every few weeks). If Claude Desktop reports authentication errors, re-run the same login command from inside the project folder:
Windows:
cd "C:\YOUR_PATH_TO_REPO\Experiment_NotebookM_MCP"
.venv\Scripts\python -m notebooklm login
Mac/Linux:
cd /YOUR_PATH_TO_REPO/Experiment_NotebookM_MCP
.venv/bin/python -m notebooklm login
Follow the same steps as the initial login:
- Sign in to Google in the browser that opens
- Wait for the NotebookLM homepage to load
- Press ENTER in the terminal to save
The storage_state.json file is overwritten with fresh cookies. You do not need to change any config or restart Claude Desktop — the updated cookies are picked up automatically on the next request.
Troubleshooting
| Problem | Fix |
|---|---|
Not authenticated or 401 Unauthorized |
Re-run notebooklm login — session expired |
Storage file not found: ...storage_state.json |
Login was not completed properly — re-run login and press ENTER in the terminal after the NotebookLM page loads |
Module not found |
Re-run .venv\Scripts\pip install -r requirements.txt inside the venv |
No module named 'pydantic_core._pydantic_core' |
Binary mismatch between Python versions — run .venv\Scripts\pip install --force-reinstall -r requirements.txt |
spawn ... python.exe ENOENT |
The .venv folder is missing — re-run python -m venv .venv then pip install -r requirements.txt |
playwright: command not found |
Run pip install playwright then playwright install chromium |
| Tools not showing in Claude Desktop | Verify all paths in the config are correct, then fully quit (system tray) and reopen Claude Desktop |
| Config changes not taking effect | Claude Desktop must be fully quit from the system tray — closing the window is not enough |
PermissionError on storage path |
NOTEBOOKLM_STORAGE_PATH must point to the .notebooklm folder, not a file |
NOTEBOOKLM_STORAGE_PATH mismatch |
Ensure the path in the config matches the default login storage: C:\Users\YOUR_USERNAME\.notebooklm |
Tool Reference
All 56 tools follow the naming convention notebooklm_{action}_{resource}.
| Category | Tools |
|---|---|
| Notebooks | notebooklm_list_notebooks, notebooklm_create_notebook, notebooklm_get_notebook, notebooklm_rename_notebook, notebooklm_delete_notebook, notebooklm_get_notebook_summary, notebooklm_get_notebook_description |
| Sources | notebooklm_list_sources, notebooklm_add_url_source, notebooklm_add_text_source, notebooklm_add_file_source, notebooklm_add_drive_source, notebooklm_get_source, notebooklm_get_source_fulltext, notebooklm_get_source_guide, notebooklm_delete_source, notebooklm_refresh_source, notebooklm_rename_source |
| Chat | notebooklm_chat_ask, notebooklm_chat_configure, notebooklm_chat_set_mode, notebooklm_chat_history |
| Artifacts | notebooklm_generate_audio, notebooklm_generate_video, notebooklm_generate_slide_deck, notebooklm_generate_quiz, notebooklm_generate_flashcards, notebooklm_generate_report, notebooklm_generate_infographic, notebooklm_generate_mind_map, notebooklm_generate_study_guide, notebooklm_generate_data_table |
| Artifact Mgmt | notebooklm_list_artifacts, notebooklm_poll_artifact, notebooklm_wait_artifact, notebooklm_download_artifact, notebooklm_delete_artifact, notebooklm_rename_artifact, notebooklm_revise_slide, notebooklm_suggest_reports |
| Notes | notebooklm_create_note, notebooklm_list_notes, notebooklm_update_note, notebooklm_delete_note |
| Sharing | notebooklm_get_share_status, notebooklm_set_public, notebooklm_share_add_user, notebooklm_share_remove_user, notebooklm_set_view_level, notebooklm_get_share_url |
| Research | notebooklm_start_research, notebooklm_poll_research, notebooklm_import_research_sources |
| Settings | notebooklm_get_output_language, notebooklm_set_output_language |
| Export | notebooklm_export |
Notes
- Uses the unofficial
notebooklm-pylibrary which reverse-engineers NotebookLM's internal APIs. Google may change these APIs at any time. - Not affiliated with or endorsed by Google.
- Session cookies expire periodically and require re-login via
notebooklm login. - Rate limits apply — the server returns error messages if you hit them.
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.