webfetch
Fetch any web page and convert it to clean, AI-ready Markdown.
Quick start
webfetch https://example.com
webfetch https://example.com -o output.md
Usage
webfetch <url> [options]
Options:
-m, --method <auto|ai|browser> Conversion method (default: auto)
-i, --images Retain images in output
-o, --output <file> Write to file instead of stdout
-t, --tokens Show token count header
-r, --rate-limit Show rate limit remaining
-h, --help Show help
Methods
- auto (default) - Try fastest method first, fallback automatically
- ai - Use Workers AI (good for static HTML)
- browser - Use headless browser (for JS-heavy pages, adds ~1-2s latency)
Examples
# Basic fetch
webfetch https://news.ycombinator.com
# JS-heavy site with browser rendering
webfetch https://spa-react-app.com -m browser
# Save to file with metadata
webfetch https://example.com -o article.md -t -r
# Include images
webfetch https://example.com -i
Rate limits
500 requests per day per IP. Check x-rate-limit-remaining header with -r flag.
Script location
<skill-dir>/scripts/webfetch.sh