Curl.Md

Curl.Md fills a specific niche: getting clean, token-minimized markdown from URLs for AI agents. It's not a general-purpose scraper—if you need rendered screenshots or deep interaction with JavaScript-heavy SPAs, look elsewhere. But for the majority of web content (blogs, docs, articles), it works well and saves real money on token costs. The browser prefix is the easiest onboarding we've seen—just prepend `curl.md/` to any URL. The CLI and SDK are solid for scripting, and the official agent plugins (Amp, Claude, OpenCode, Pi) make integration trivial. Where it bites: rate limits on the free tier can be restrictive for batch processing, and JS-heavy or dynamically loaded content may not render fully (the tool fetches HTML, not a browser-rendered page). Also, there's no dedicated enterprise support—you're relying on open-source community and docs. Compared to alternatives like Jina Reader or Firecrawl, Curl.Md is more token-conscious and lighter weight, but Firecrawl offers better JS rendering and pagination. If you're a solo dev or small team building agent workflows and want to cut costs, Curl.Md is a solid choice. For high-volume production use with complex sites, consider a paid scraping API.