notebooklm-integration（NotebookLM 集成）

NotebookLM Integration Skill

This skill enables you to leverage the full power of Google NotebookLM through the unofficial notebooklm-py Python library. It provides programmatic access to features that aren't available in the web UI, including batch operations, custom format exports, and advanced automation capabilities.

When to Use This Skill

Use this skill when you need to:

• Create, list, rename, or delete NotebookLM notebooks
• Import various source types (URLs, YouTube videos, PDFs, text files, Google Drive, etc.)
• Ask questions and chat with your notebooks using custom personas
• Run web and Drive research agents with auto-import capabilities
• Generate Audio Overviews (podcasts) in multiple formats and languages
• Create Video Overviews with different visual styles
• Generate Slide Decks (PDF/PPTX) and Infographics (PNG)
• Create Quizzes and Flashcards in multiple formats (JSON, Markdown, HTML)
• Download all generated artifacts locally or export to Google Docs/Sheets
• Share notebooks with specific permissions and view level controls

Quick Start

Installation

First, ensure you have the notebooklm-py library installed:

pip install notebooklm-py

Basic Usage Patterns

Python API

from notebooklm import NotebookLMClient

# Initialize client
client = NotebookLMClient()

# Create a new notebook
notebook = client.create_notebook("My Research Project")

# Add sources
notebook.add_source(url="https://example.com/research-paper.pdf")
notebook.add_source(youtube_url="https://youtube.com/watch?v=abc123")
notebook.add_source(file_path="./documents/report.txt")

# Ask questions
response = notebook.ask("What are the main findings in these sources?")
print(response.text)

# Generate audio overview
audio = notebook.generate_audio_overview(
    format="deep-dive",
    length="medium",
    language="en"
)
audio.save("./outputs/podcast.mp3")

CLI Usage

# Create notebook
notebooklm notebook create "My Research"

# Add sources
notebooklm notebook add-source "My Research" --url https://example.com/paper.pdf
notebooklm notebook add-source "My Research" --youtube https://youtube.com/watch?v=abc123

# Ask questions
notebooklm notebook ask "My Research" "Summarize the key points"

# Generate content
notebooklm notebook audio "My Research" --format deep-dive --length medium
notebooklm notebook video "My Research" --style cinematic
notebooklm notebook slide "My Research" --format detailed

# Download artifacts
notebooklm notebook download "My Research" --format mp3 --output ./podcasts/

Advanced Features

Research Automation

# Run web research with auto-import
research_notebook = client.research_web(
    query="latest developments in quantum computing",
    max_sources=10,
    mode="deep"  # or "fast"
)

# Run Drive research
drive_notebook = client.research_drive(
    folder_id="your-drive-folder-id",
    query="machine learning papers"
)

Batch Operations

# Import multiple sources at once
sources = [
    {"type": "url", "value": "https://example1.com"},
    {"type": "youtube", "value": "https://youtube.com/watch?v=..."},
    {"type": "file", "value": "./document.pdf"}
]

notebook.add_sources(sources)

# Generate multiple content types
formats = ["mp3", "mp4", "pdf", "png"]
for fmt in formats:
    notebook.download_artifacts(format=fmt, output_dir=f"./outputs/{fmt}")

Custom Personas

# Set a custom persona for more focused responses
notebook.set_persona(
    "You are a technical expert specializing in machine learning. "
    "Provide detailed, accurate explanations with code examples when relevant."
)

Output Formats

Audio Overview

• Formats: deep-dive, brief, critique, debate
• Lengths: short, medium, long
• Languages: 50+ supported
• Output: MP3/MP4

Video Overview

• Formats: explainer, brief, cinematic
• Styles: 9 visual styles plus cinematic-video alias
• Output: MP4

Slide Deck

• Formats: detailed, presenter
• Output: PDF, PPTX

Infographic

• Orientations: 3 (portrait, square, landscape)
• Detail levels: 3 (low, medium, high)
• Output: PNG

Quiz & Flashcards

• Configurable quantity and difficulty
• Output: JSON, Markdown, HTML

Best Practices

1. Error Handling: The library uses undocumented Google APIs that may change - implement retry logic and fallback mechanisms
2. Rate Limits: Be mindful of usage quotas to avoid throttling
3. Cleanup: Temporary files are cleaned up automatically, but manage your output directories
4. Authentication: Uses your Google credentials - ensure you're logged in via browser auth flow
5. Organization: Create engagement-specific notebooks for different projects

Updating the Skill

To update this skill to the latest version from the GitHub repository, follow these steps:

1. Clone or pull the latest version of the notebooklm-py repository:

   git clone https://github.com/teng-lin/notebooklm-py.git
   # or if you already have it:
   cd notebooklm-py && git pull

1. Re-run the installation process:

   pip install -e .  # for development mode, or just pip install notebooklm-py

1. If you're using the OpenClaw skill, you can update it by re-running the skill creation process from the latest repository.

Troubleshooting

• If APIs break, check the Troubleshooting guide in the notebooklm-py repo
• For authentication issues, re-run the login process
• Rate limit errors require reducing request frequency or implementing exponential backoff
• Some features may require specific Google Workspace permissions

Related Skills

• ai-agent-development - For building agents that utilize NotebookLM capabilities
• audio-transcriber - For processing generated audio content
• video-frames - For extracting frames from video overviews
• app-builder - For creating full applications around NotebookLM workflows

Bet, Boss. This skill puts the full power of NotebookLM at your fingertips. What notebook shall we create first? 😉