finds.dev

teng-lin/notebooklm-py

Unofficial Python API and agentic skill for Google Gemini Notebook. Full programmatic access to Gemini Notebook's features—including capabilities the web UI doesn't expose—via Python, CLI, and AI agents like Claude Code, Codex, and OpenClaw.

View on GitHubHomepage ↗

★ 20,073PythonMITupdated Nov 2026

// the take

An unofficial Python/CLI/MCP client for Google's NotebookLM (now Gemini Notebook) that reverse-engineers the web and mobile APIs to expose nearly every feature of the product programmatically, including things the web UI doesn't surface like bulk artifact export. It's aimed at developers and AI agent builders who want to drive NotebookLM as a grounded research/content-generation backend from scripts, CI, or Claude Code/Codex rather than clicking through the browser.

Feature coverage is unusually complete and matches the real product surface: notebooks, sources, chat, notes, labels, research agents, sharing, and every studio artifact type (audio, video, slides, quizzes, flashcards, reports, data tables, mind maps), exposed consistently across Python API, CLI, MCP server, and a REST server. It ships two independent backends - the standard web batchexecute transport and a separately reverse-engineered Android gRPC protocol with checked-in protobuf definitions - so there's a fallback path if one breaks. The auth model is more mature than typical scraping libraries: Playwright login, browser-cookie import, and a master-token mode built specifically for headless/unattended use with self-healing cookie refresh. There's real investment in reliability engineering for a reverse-engineered client - a fault-injection harness, nightly and offline qualification workflows, VCR-style cassette scrubbing for recorded RPC tests, plus CodeQL and dependency-audit CI.

Everything depends on undocumented internal Google RPCs and a reverse-engineered mobile gRPC protocol - Google can change either without notice and owes this project nothing, which the README itself admits by calling it best for prototypes rather than production use. The master-token and browser-cookie auth paths mean the library holds full-account credentials rather than a scoped API key; the docs explicitly warn the master token is a powerful full-account credential to guard carefully, which is a much bigger blast radius if a local profile store leaks than a normal API wrapper would have. There's no official rate limit or SLA to code against, so scheduled or heavy use (the README's own cron-based audio-briefing recipe, for instance) is exposed to Google's undocumented throttling or anti-bot detection with no visibility into what triggers it. The default install path also pulls in Playwright plus a ~170MB Chromium download for anyone not using cookie import or master token, which is a heavy dependency footprint for what's conceptually a thin API client.

We read the README, the directory tree and a key source file in November 2026. Nobody pays to be listed. How we judge →

// at a glance

Stars
20,073
Forks
2,671
Language
Python
License
MIT
Last commit
1 Nov 2026

// nearby

If notebooklm-py isn't quite it.

All repos →