Skip to content

notebooklm-py

by Teng LinConnectorResearch something

Lets your AI agent use Gemini Notebook to make podcasts, quizzes and slides from sources

View on GitHub

About notebooklm-py

An unofficial connector for Google's Gemini Notebook (formerly NotebookLM). Sign in with Google and your agent can add sources, ask questions with cited answers, and make podcasts, slides and quizzes.

For StudentsResearchersNotebookLM users

Made by Teng Lin

An open-source project by developer Teng Lin, published on GitHub. It is unofficial and not made by Google.

Stars
★ 20k
License
MIT
Cost
Free
Level
Intermediate
Runs
Local

Overview

notebooklm-py is an unofficial, open-source connector for Google's Gemini Notebook, formerly NotebookLM; it is not made by Google. Once you sign in with your Google account, your agent can create notebooks, add web pages, PDFs or YouTube videos as sources, ask questions with cited answers, and generate audio overviews, slide decks, quizzes, flashcards and mind maps, then download them as files. It also works as a command-line tool and an agent skill. It suits students, researchers and writers who already use NotebookLM.

It can download quizzes, flashcards and mind maps as files in bulk, something the NotebookLM web app itself does not offer.

Try these first

Start with the first one: it is a small task you can check easily.

  1. List my notebooks and the sources in [notebook name]. Don't create or change anything.First try · read-only

  2. Create a notebook called [topic], add these links as sources: [paste links], and tell me the key themes.writes to your account

  3. Generate a 10-question quiz from that notebook and save it as a Markdown file.writes to your account

  4. Make an audio overview from the notebook and download it as an audio file.writes to your account

Example prompts written by Agent Flows Hub.

Capabilities

From the creator's documentation.

ToolsetWhat your agent can doAccess
NotebooksCreate, copy, rename, list and delete notebooksRead + write
SourcesAdd web pages, YouTube videos, PDFs, Word files, Google Drive files or pasted textRead + write
ChatAsk questions and get answers with citations from your sourcesRead
Generated contentAudio and video overviews, slide decks, quizzes, flashcards, infographics, reports and mind mapsRead + write
DownloadsSave outputs as MP3, MP4, PDF, PPTX, PNG, CSV, JSON or Markdown filesRead
Research and sharingRun web research that imports sources, and manage sharing linksRead + write

Good to know

You have a pile of articles, PDFs and videos to study and want NotebookLM's summaries, quizzes and audio overviews without clicking through the site. notebooklm-py lets your agent add the sources, ask questions and save the outputs as files.

Limits

  • It relies on undocumented Google endpoints, so features can stop working when Google changes them, and heavy use may be throttled.
  • Sign-in saves your Google session on your computer and the connector acts with your account. The MCP server is marked experimental, so tool names can change between releases.

Notes by (original project name: notebooklm-py), written from the creator's documentation, not from running it ourselves. notebooklm-py README · notebooklm-py MCP server guide

Stuck setting it up?

Tell us what happened. We take a look first; if it needs paid work, you get an estimate before anything starts.

Get help

Other connectors

  • Notion MCP Lets your AI agent search, read and edit pages in your Notion workspace
  • Open Computer Use Lets your AI agent see and operate the apps on your computer by clicking and typing
  • Pascal Editor Lets your AI agent draw rooms, walls and furniture in a 3D home model you can then edit
  • PayPal MCP Lets your AI assistant create and send PayPal invoices, look up orders and review disputes
  • Pipeboard Meta Ads MCP Lets your AI assistant check, create and adjust your Facebook and Instagram ad campaigns
  • Playwright MCP Lets your AI agent open web pages, click, type and fill in forms in a real browser

See all 52 connectors · An MCP server is missing or disconnected

GitHub