WebPeel — Your AI Can Finally Read Any Website

# WebPeel — The Web Data Platform for AI Agents > Fetch, search, crawl, extract, monitor, screenshot, and research any URL. One API call. 65-98% token savings. ## Quick Decision Tree Choose the right approach: - Static page → `webpeel <url>` (default, fastest) - JavaScript SPA (React, Vue, Next.js) → `webpeel <url> --render` - Bot-protected site (Cloudflare) → `webpeel <url> --stealth` - Infinite scroll / lazy content → `webpeel <url> --render --action 'scroll:bottom' --action 'wait:2000'` - Interact (click, type) → `webpeel <url> --render --action 'click:.btn' --action 'wait:1000'` - Screenshot → `webpeel screenshot <url>` - YouTube transcript → `webpeel <youtube-url>` - PDF → `webpeel <pdf-url>` - Structured data → `webpeel <url> --schema product --json` - Web search → `webpeel search "query"` - Monitor for changes → `webpeel watch <url>` ## Capabilities - **Fetch**: Extract clean markdown from any URL. Auto-handles JS rendering, Cloudflare, CAPTCHAs. - **Search**: Web search with structured results. Smart search detects intent (restaurants, products, flights, hotels). - **Crawl**: Crawl sites up to 100 pages with dedup and rate limiting. - **Extract**: JSON schema extraction. Built-in schemas: product, article, recipe, job, event, contact, review, listing. - **Monitor**: Track pages for changes with diffs and CSS selector targeting. - **Screenshot**: Full-page screenshots with device emulation. Visual diff comparison. - **Research**: Autonomous research agent — plans queries, fetches sources, synthesizes with citations. - **YouTube**: Full transcripts with timestamps, 50+ languages. - **PDF**: Text extraction from PDFs. - **Act**: Browser automation — click, fill forms, scroll, interact, then extract. ## Browser Actions (--action flag) For dynamic pages. Auto-enables browser rendering. Chain multiple: - `scroll:bottom` — scroll to bottom (infinite scroll) - `scroll:down:500` — scroll down 500px - `wait:2000` — wait 2 seco