notebooklm

Query NotebookLM notebooks for source-grounded, citation-backed answers.

103|30|Updated Dec 24, 2025
One-click install
npx skills add https://github.com/CommandCodeAI/agent-skills --skill notebooklm-commandcodeai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: notebooklm
Source: https://github.com/CommandCodeAI/agent-skills/tree/main/skills/notebooklm-skill
Command: npx skills add https://github.com/CommandCodeAI/agent-skills --skill notebooklm-commandcodeai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires patchright, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This skill enables you to query Google NotebookLM notebooks directly from Claude Code and receive source-grounded, citation-backed answers, reducing hallucinations by restricting responses to your uploaded documents.

Core Features & Use Cases

  • Browser automation: Launches a fresh NotebookLM session per query to fetch precise information from your documents.
  • Library management: Manage NotebookLM notebooks with metadata, topics, and active selections to organize your docs.
  • Persistent authentication support: Hybrid authentication using a browser profile and manual cookie injection to maintain sessions across runs.
  • Use Case: When you need an answer grounded in your own documents, e.g., querying a technical notebook and getting citations from your materials.

Quick Start

Start by authenticating and adding a notebook:

  • python scripts/run.py auth_manager.py setup
  • python scripts/run.py notebook_manager.py add --url "https://notebooklm.google.com/notebook/..." --name "Descriptive Name" --description "Notebook contents" --topics "topic1,topic2" Then ask a question:
  • python scripts/run.py ask_question.py --question "What does this notebook contain?" --notebook-id notebook-id

Frequently Asked Questions about notebooklm

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get citation-backed answers from my NotebookLM documents using Claude Code?

To get citation-backed answers, this skill queries NotebookLM notebooks via Claude Code using a stateless question model, restricting responses to your uploaded documents to reduce hallucinations and provide source-grounded results.

How does browser automation handle NotebookLM authentication for automated queries?

NotebookLM authentication uses a hybrid workflow combining a persistent browser profile with manual cookie injection via a state.json file, maintaining browser sessions across multiple automated query runs.

Can I manage multiple NotebookLM notebooks with metadata and topics in Claude Code?

You can manage NotebookLM notebooks by adding them with descriptive names, descriptions, and topics using the notebook manager script, organizing your document library for targeted querying.

Do I need the patchright library to run NotebookLM queries in Claude Code?

Yes, patchright is required for NotebookLM queries because it provides the browser automation and cookie handling needed to launch fresh sessions and fetch document data.

Why are my NotebookLM automated queries returning session errors?

Session errors usually occur because hybrid authentication is not properly configured; run the auth manager setup script and ensure manual cookies are correctly injected into the state.json file.