Is your site readable
by AI agents?
Paste any URL and check its AI-search readiness: llms.txt, Markdown for agents, structured data, sitemap, and crawler rules — each gap explained with how to fix it.
How do I check if my site is readable by AI agents? Paste your URL below. The checker inspects robots.txt for AI crawler rules, the presence of llms.txt and llms-full.txt, whether the server returns Markdown when asked (Accept: text/markdown), JSON-LD structured data, the XML sitemap, and per-page Markdown link relations — then scores your setup out of 9.
AI agent readiness checker
Checks run from our server against public files of the URL you enter.
Unlike our other tools, this checker needs a small server-side fetch to inspect the public files of the URL you enter (browsers block cross-site reads for security). We don't store URLs or results.
What does “AI agent readiness” mean?
It is how easily AI assistants and agents (ChatGPT search, Claude, Perplexity, Grok, coding agents) can read, trust, and cite your pages. The basics: allow AI crawlers, describe your site with llms.txt, serve clean Markdown, and expose structured data so machines understand what each page is.
What is Markdown for agents?
AI agents increasingly request pages with Accept: text/markdown so they get clean text instead of full HTML. If your server (or a small Worker) returns Markdown for that request, agents spend fewer tokens and can quote your content more accurately. This site does it — try curl -H "Accept: text/markdown" https://toolpantry.app/.
Why does this checker need a server while your other tools run locally?
Browsers block one website from reading another site's files (CORS). To inspect robots.txt or llms.txt on your domain, something server-side has to fetch them. We only fetch the public files of the URL you enter, and we don't store the URL or the results.
How do I fix the failed checks?
The open-source kit on GitHub walks through every item: generating Markdown versions of your pages, publishing llms.txt, adding link relations, and wiring content negotiation on a Cloudflare Worker — free-plan friendly. Each check above fails with a specific reason so you know exactly what to add.
Is a perfect score required?
No. Crawler access and a sitemap are table stakes; llms.txt, Markdown versions, and JSON-LD are where AI visibility actually improves. Even 4–6 out of 9 puts you ahead of most sites today.