Gemini Notebook CLI & MCP Server
Programmatic access to Gemini Notebook (formerly NotebookLM) via CLI and MCP: notebooks, sources, podcasts, video, reports and research from an agent
High risk
We rate an entry high when the tool writes to external systems, handles money, production databases or secrets, or runs arbitrary commands. The CLI installs it only with your consent.
Why this level
- Manages data in the Gemini Notebook account: creates and deletes notebooks and publishes share links
- Stores and uses Google session cookies to access the account
Install
Manual install
uv tool install notebooklm-mcp-cliInstalls the nlm command and the notebooklm-mcp server binary. Alternatives via uvx, pip and pipx exist. After install run nlm login.
This is third-party code. Review the repository files before installing.
What it does
The project gives an agent access to Gemini Notebook two ways: the nlm terminal command and the notebooklm-mcp MCP server for clients such as Claude Code, Claude Desktop, Gemini CLI and Cursor. From an agent you can create notebooks, add sources (links, YouTube, text, files, Google Drive documents), ask questions against sources, run web research and generate outputs: podcasts, video, reports, quizzes, flashcards, mind maps, slides and infographics. It also has batch operations, cross-notebook queries, multi-step pipelines and a plan-usage check. Sign-in goes through nlm login, which pulls cookies from a managed browser session, so no separate API key is needed. Downloading artifacts, sharing notebooks and deleting them are available from the same commands.
Who it is for. For researchers, writers and analysts who work in Gemini Notebook and want to drive it from an agent or scripts.
Good fit when
- You want to assemble a notebook from sources and query it from an agent
- You want to generate a podcast, report or video from a notebook's material
- You want to automate Gemini Notebook routines with a script or pipeline
Not a fit when
- You have no access to Gemini Notebook or a Google account to sign in
- You need an official, guaranteed API rather than access through a browser session
Example request
Create a notebook from these three links and turn it into a 10-minute podcastLimitations
This is an unofficial client: access relies on Google account cookies, so changes to Google's web app can temporarily break commands. You need a working account with Gemini Notebook access, and enterprise account support is marked experimental. The Google service is not reachable from Russia without a VPN. Some operations change account data, including creating and deleting notebooks and publishing links, so deletion and generation are behind confirmation.
How to disable. Remove the MCP server from the agent config (for example claude mcp remove gemini-notebook-mcp) and uninstall the tool: uv tool uninstall notebooklm-mcp-cli. The login cache lives in ~/.notebooklm-mcp-cli.
MCP
- Transport
- stdio
- Authentication
- OAuth
Security check
- Manages data in the Gemini Notebook account: creates and deletes notebooks and publishes share links
- Stores and uses Google session cookies to access the account
README in short
The README describes two interfaces to Gemini Notebook: the nlm terminal command and the notebooklm-mcp MCP server, both installed from the single notebooklm-mcp-cli package. It covers client setup via nlm setup add, sign-in via nlm login with cookie extraction, and a capability table: sources, queries, studio outputs, research, sharing, batch and cross-notebook operations, pipelines and a plan check. It adds automatic session recovery, an unattended mode via nlm auth refresh and a desktop extension. The package is published on PyPI under the MIT license.
FAQ
Do I need an API key?
No. Access goes through a Google sign-in via nlm login, which stores session cookies locally.
Is this an official Google tool?
No. It is an unofficial project that works through a browser session rather than a public Gemini Notebook API.
Related
A self-hosted knowledge base with block-level references and a built-in MCP server for connecting AI agents to your notes
A CLI for every Google Workspace API with JSON output and agent skills: Drive, Gmail, Calendar, Sheets and more
Local search over Markdown notes, docs and meeting transcripts: keywords, semantic search and reranking, with an MCP server
A task manager for AI-driven development: breaks a PRD into dependent tasks and guides the agent through them via MCP or CLI