---
name: notebooklm-kd
description: Automate NotebookLM research workflow - from YouTube links to structured reports (Knowledge Distillery).
homepage: https://notebooklm.google.com/
metadata:
  {
    "openclaw":
      {
        "emoji": "🎓",
        "requires": { "bins": ["uv"] },
        "install":
          [
            {
              "id": "uv-brew",
              "kind": "brew",
              "formula": "uv",
              "bins": ["uv"],
              "label": "Install uv (brew)",
            },
          ],
      },
  }
---

# NotebookLM-KD (Knowledge Distillery)

Automate the workflow of extracting structured knowledge from YouTube videos or articles using NotebookLM + Gemini AI.

## Quick Start

Generate a research report from YouTube videos:

```bash
uv run {baseDir}/scripts/kd.py research \
  --topic "Financial Research" \
  --urls "https://youtube.com/watch?v=xxx,https://youtube.com/watch?v=yyy" \
  --format briefing
```

## Parameters

- `--topic`: Topic name (used for folder name)
- `--urls`: Comma-separated YouTube/article URLs
- `--format`: Report format (briefing/study-guide/blog)
- `--output`: Output path (default: workspace/{topic}_KD/)

## Output

Creates a complete research folder:

```
workspace/{topic}_KD/
├── README.md                          # Index with NotebookLM link
└── YYYY-MM-DD_{topic}_KD報告.md        # Full report
```

## Notes

- Uses OpenClaw browser automation (no API key needed)
- Requires Google account with NotebookLM access
- Max 50 sources per NotebookLM notebook
- Public YouTube videos only (with transcripts)
- Report generated by Gemini AI via NotebookLM

## Workflow

1. Opens NotebookLM in browser
2. Creates new notebook
3. Batch imports URLs
4. Waits for Gemini analysis
5. Generates report (selected format)
6. Copies report content
7. Creates folder structure
8. Saves report with templates
9. Records NotebookLM link

Total time: ~3-5 minutes for 4 videos