🎯 Multi-Source Content → NotebookLM Smart Processor

One sentence turns into a podcast, PPT, mind map, quiz...

🔀 Fork of joeseesun/anything-to-notebooklm with expanded features and improvements

Quick Start • Supported Formats • Usage Examples • FAQ

---

✨ What Is This?

A Claude Code Skill that lets you use natural language to turn any content into any format.

You say: Turn this article into a podcast
AI:      ✅ 8-minute podcast generated → podcast.mp3

You say: Convert this EPUB ebook into a mind map
AI:      ✅ Mind map generated → mindmap.json

You say: Make this YouTube video into slides
AI:      ✅ 25-page PPT generated → slides.pdf

How it works: Automatically fetch content from multiple sources → Upload to Google NotebookLM → AI generates your desired format

🚀 Supported Content Sources (15+ Formats)

📱 Social Media YouTube Videos (auto subtitle extraction) 🌐 Web Any web page (news, blogs, docs) Search keywords (auto-aggregate results) 📄 Office Documents Word (.docx) PowerPoint (.pptx) Excel (.xlsx)

📚 Ebooks & Documents PDF (including scanned with OCR) EPUB (ebooks) Markdown (.md) 🖼️ Images & Audio Images (JPEG/PNG/GIF, auto OCR) Audio (WAV/MP3, auto transcription) 📊 Structured Data CSV/JSON/XML ZIP archives (batch processing)

Powered by: Microsoft markitdown

🎨 What Can It Generate?

Output Format	Use Case	Gen Time	Trigger Examples
🎙️ Podcast	Listen on the go	2–5 min	"generate podcast", "make audio"
📊 Slides	Team presentations	1–3 min	"make PPT", "generate slides"
🗺️ Mind Map	Visualize structure	1–2 min	"draw a mind map", "concept map"
📝 Quiz	Self-assessment	1–2 min	"generate quiz", "create test"
🎬 Video	Visual content	3–8 min	"make a video"
📄 Report	Deep analysis	2–4 min	"generate report", "write summary"
📈 Infographic	Data visualization	2–3 min	"make infographic"
📋 Flashcards	Memory reinforcement	1–2 min	"make flashcards"
📊 Data Table	Structured comparison	1–2 min	"create data table"

Fully natural language — no commands to memorize!

⚡ Quick Start

Prerequisites

• ✅ uv (universal Python tool runner)
• ✅ Git (pre-installed on macOS/Linux)

That's it! All other dependencies are auto-installed.

Installation (3 Steps)

# 1. Clone to Claude skills directory
cd ~/.claude/skills/
git clone https://github.com/stanvx/anything-to-notebooklm
cd anything-to-notebooklm

# 2. One-click install all dependencies
./install.sh

install.sh creates a local .venv with all dependencies and convenience wrappers in ./bin/. Optionally add to PATH: export PATH="$HOME/.claude/skills/anything-to-notebooklm/bin:$PATH"

First Use

# NotebookLM authentication (one-time only)
./bin/notebooklm login
./bin/notebooklm list  # Verify success

# Environment check (optional)
./check_env.py

OpenClaw (AgentSkills) Setup

OpenClaw loads skills from <workspace>/skills/ or ~/.openclaw/skills/. If you haven’t bootstrapped OpenClaw yet, run openclaw setup first.

mkdir -p ~/.openclaw/workspace/skills
git clone https://github.com/stanvx/anything-to-notebooklm \
  ~/.openclaw/workspace/skills/anything-to-notebooklm
cd ~/.openclaw/workspace/skills/anything-to-notebooklm
./install.sh

💡 Usage Examples

Scenario 1: Team Sharing — Ebook → PPT

You: Turn this book into slides /Users/joe/Books/sapiens.epub

AI automatically:
  ✓ Extracts ebook content (150K words)
  ✓ AI distills key insights
  ✓ Generates professional slides

✅ Result: /tmp/sapiens_slides.pdf (25 pages, 3.8 MB)
💡 Use: Ready for book club presentation

Scenario 2: Self-Assessment — Video → Quiz

You: Generate a quiz from this YouTube video https://youtube.com/watch?v=abc

AI automatically:
  ✓ Extracts video subtitles
  ✓ AI analyzes key knowledge points
  ✓ Auto-generates questions

✅ Result: /tmp/video_quiz.md (15 questions: 10 multiple choice + 5 short answer)
💡 Use: Test your understanding

Scenario 3: Information Synthesis — Multi-Source → Report

You: Combine these into a report:
    - https://example.com/article1
    - https://youtube.com/watch?v=xyz
    - /Users/joe/research.pdf

AI automatically:
  ✓ Aggregates 3 different sources
  ✓ AI synthesizes and analyzes
  ✓ Generates comprehensive report

✅ Result: /tmp/multi_source_report.md (7 chapters, 15.2 KB)
💡 Use: Complete topic research report

Scenario 4: Document Digitization — Scan → Text

You: Convert this scanned image to a document /Users/joe/scan.jpg

AI automatically:
  ✓ OCR recognizes text in image
  ✓ Extracts as plain text
  ✓ Generates structured document

✅ Result: /tmp/scan_document.txt (95%+ accuracy)
💡 Use: Digital archiving of scanned documents

🎯 Core Features

🧠 Smart Recognition

Auto-detects input type — no manual specification needed

https://youtube.com/watch?v=xxx  → YouTube Video
/path/to/file.epub               → EPUB Ebook
"search 'AI trends'"             → Search Query

🚀 Fully Automated

From acquisition to generation — seamless end-to-end

Input → Fetch → Convert → Upload → Generate → Download
       ︿_________ Fully Automated _________︿

🌐 Multi-Source Integration

Mix and match multiple content sources

Article + Video + PDF + Search Results → Comprehensive Report

🌍 50+ Languages

Generate content in any of 80+ supported languages

notebooklm language set es    # Spanish podcasts
notebooklm language set ja    # Japanese reports

🔬 AI Research Agent

Auto-discover and import sources

notebooklm source add-research "quantum computing" --mode deep --import-all

📦 Architecture

┌─────────────────────────────────────┐
│       User Natural Language Input    │
│  "Turn this article into a podcast   │
│   https://..."                       │
└──────────────┬──────────────────────┘
               │
               ▼
┌─────────────────────────────────────┐
│        Claude Code Skill             │
│  • Smart source type detection       │
│  • Auto-invoke appropriate tools     │
└──────────────┬──────────────────────┘
               │
               │
               ▼
┌─────────────────────────────────────┐
│         markitdown                   │
│  • Convert files to text             │
└──────────────┬──────────────────────┘
               │
               ▼
│         NotebookLM API               │
│  • Upload content sources            │
│  • AI-generate target formats        │
└──────────────┬──────────────────────┘
               │
               ▼
┌─────────────────────────────────────┐
│         Generated Files              │
│  .mp3 / .mp4 / .pdf / .json / .md   │
│  .png / .csv / .html                 │
└─────────────────────────────────────┘

🔧 Advanced Usage

Specify Existing Notebook

Add this article to my "AI Research" notebook https://example.com

Batch Processing

Generate podcasts from all of these:
2. https://example.com/article2
3. /Users/joe/notes.md

ZIP Batch Conversion

Turn all documents in this archive into a podcast /path/to/files.zip

Auto-extract, identify, convert, and merge.

Research → Content Pipeline

# Discover sources automatically, then generate
notebooklm source add-research "climate change policy" --mode deep --import-all
notebooklm generate audio "Focus on policy solutions" --format debate --wait
notebooklm download audio ./climate-podcast.mp3

Chat with Your Sources

notebooklm ask "What are the key themes across all sources?"
notebooklm ask "Compare the two viewpoints" -s src1 -s src2

🐛 Troubleshooting

NotebookLM Authentication Failed

./bin/notebooklm auth check --test  # Full diagnostic
./bin/notebooklm login              # Re-authenticate
./bin/notebooklm list               # Verify

Environment Check

./check_env.py       # 5-point comprehensive check
./install.sh         # Reinstall

🤝 Contributing

PRs, Issues, and suggestions welcome!

❓ FAQ

Q: What languages are supported?

A: NotebookLM supports 80+ languages. Set with notebooklm language set <code>. English and Chinese work best.

Q: Whose voice is the podcast?

A: Google AI voice synthesis. English features two AI hosts in dialogue; other languages typically use single narrator.

Q: What are the content length limits?

A:

• Minimum: ~500 words
• Maximum: ~500K words
• Recommended: 1,000–10,000 words for best results

Q: Can I use this commercially?

A:

• This Skill: MIT open source, free to use
• Generated content: Subject to NotebookLM Terms of Service
• Source content: Subject to original content copyright
• Recommendation: For personal learning and research only

Q: What podcast formats are available?

A: Four formats via --format: deep-dive (default, thorough exploration), brief (concise overview), critique (critical analysis), and debate (two-sided discussion). Three lengths: short, default, long.

Q: What video styles are available?

A: Nine styles via --style: auto, classic, whiteboard, kawaii, anime, watercolor, retro-print, heritage, paper-craft.

📄 License

MIT License

🙏 Acknowledgements

• Google NotebookLM — AI content generation
• Microsoft markitdown — File format conversion
• notebooklm-py — NotebookLM CLI

📮 Contact

• Issues: GitHub Issues
• Original Repo: joeseesun/anything-to-notebooklm

---

If you find this useful, please give it a ⭐ Star!

Fork of joeseesun/anything-to-notebooklm