Summarize
A versatile content processing skill sourced from ClawHub (151k+ downloads, 578 stars, by steipete). Takes any URL or file path and produces a concise summary regardless of format.
88Trust High
by steipeteresearchbeginnerv1.0.0updated Mar 7, 2026
54.2kTotal Runs
92.0%Success Rate
3.3kInstalls
88Trust Score
Tags
#summarization#pdf#youtube#audio#web#content#extraction#clawhub
Required Tools
bashweb_fetchInputs
| Name | Type | Description | Req |
|---|---|---|---|
| source | text | URL or file path to summarize. Supports: web URLs, PDF files, YouTube links, audio files, image files. | yes |
| length | text | Summary length: "brief" (1-2 sentences), "standard" (1-2 paragraphs, default), "detailed" (comprehensive). | -- |
| format | text | Output format: "text" (default), "bullets", "json". | -- |
Outputs
| Name | Type | Description | Req |
|---|---|---|---|
| summary | json | JSON with fields: title, summary_text, key_points (array), source_type (web|pdf|youtube|audio|image), word_count, source_url. | yes |
Compatible Skills
SKILL.md
---
name: summarize
description: Summarize any URL or file — web pages, PDFs, images, audio, and YouTube videos. Multi-format content extraction and summarization.
---
# Summarize
Summarize any content from URLs or files.
## Quick Start
### Web Pages
```bash
# Fetch and summarize
curl -s "URL" | html2text | head -5000 | summarize
# Using readability
curl -s "URL" | readability | summarize
```
### PDFs
```bash
# Extract text and summarize
pdftotext document.pdf - | summarize
# Using markitdown
uvx markitdown document.pdf | summarize
```
### YouTube Videos
```bash
# Download transcript
yt-dlp --write-auto-sub --skip-download "https://youtube.com/watch?v=VIDEO_ID"
# Summarize transcript
cat video.en.vtt | summarize
```
### Audio Files
```bash
# Transcribe with whisper
whisper audio.mp3 --model base
# Summarize transcript
cat audio.txt | summarize
```
## Length Options
| Option | Description |
|--------|-------------|
| brief | 1-2 sentences |
| standard | 1-2 paragraphs (default) |
| detailed | Comprehensive summary |
## Output Formats
### Text (default)
Plain text summary
### Bullets
Key points as bullet list
### JSON
```json
{
"title": "Content Title",
"summary_text": "Summary...",
"key_points": ["Point 1", "Point 2"],
"source_type": "web",
"word_count": 500,
"source_url": "https://..."
}
```
## Supported Sources
- Web pages (HTML)
- PDF documents
- YouTube videos (via transcript)
- Audio files (via transcription)
- Images (via OCR)
- Text files
- Office documents (.docx, .pptx)
## Example Workflow
1. Extract content from source
2. Clean and normalize text
3. Generate summary at specified length
4. Format output as requested
## Best Practices
- Use brief for quick overviews
- Use standard for most content
- Use detailed for research/analysis
- JSON format for programmatic use
- Check source_type for verification