Agent readiness check
Paste any page URL and get a 0-100 readiness score for AI agents: structured data, crawl access, machine readability, and agent guidance (llms.txt, sitemap) - with the specific gaps and the fix for each.
AI agents can only cite what they can fetch and parse. Readiness is three layers: access (robots.txt plus CDN and WAF rules), extraction (server-rendered HTML, clear headings, structured data - most AI crawlers do not execute JavaScript), and guidance (an accurate sitemap and llms.txt). This check scores one URL across those layers and names the specific gap in each.
Free and anonymous. Rate-limited to 20 checks per hour per IP. We fetch the page, /robots.txt, and /llms.txt.
Full report (JSON)
Readiness measures whether agents CAN read you; it does not measure whether they cite you. Next: check sitewide crawler policy with the crawler policy check, or run the full free AI visibility audit.
Why can't ChatGPT read my website?
There are exactly three places the failure can live. Access: the crawler is turned away by robots.txt, a CDN rule, or a WAF default. Extraction: the crawler gets in but the content lives in client-side JavaScript, unlabeled markup, or buried paragraphs. Guidance: nothing tells agents which pages are canonical.
Each layer has a different fix, which is why a single score without a breakdown is not actionable - and why this check reports the layer, not just the number.
Do AI crawlers execute JavaScript?
Mostly no. Measurements by Vercel and MERJ found the large majority of AI crawler fetches cannot run JavaScript, so a client-only single-page app reads as an empty page to them regardless of what a human sees.
Server-side rendering or prerendering puts your content into the initial HTML, where every bot - and every impatient human - can read it.
This score says whether agents CAN read a page. The free audit asks five engines whether they DO cite you - prompt by prompt - and classifies every miss by the layer that caused it.
What is llms.txt and do I need one?
It is a plain-text file at your site root that lists your canonical, high-value pages for AI agents - a sitemap written for language models. Ahrefs measured adoption at roughly one in ten domains, and no engine documents it as required for citation.
It is, however, a low-effort guidance signal with no downside: pair it with an accurate XML sitemap and keep both pointed at the pages you most want represented.
What makes a page extractable?
Answer-first sections - a direct answer in the first sentence under each heading - one concept per section, lists for parallel facts, and schema that mirrors the visible text. Engines parse by section, not by page, so every important fact needs its own structural home.
Extraction failures are the quietest gap class: the bot got in, saw noise, and cited someone clearer.
See what an extractable, front-loaded answer block looks like inside a real package: open the sample fix.
What this check does not do
It measures capability at one URL. It cannot see the rest of your site, and it cannot see outcomes.
The free audit runs your buyer prompts across five engines sitewide and shows whether readable ever became cited - and when it did not, which layer to fix first.
Frequently asked questions
Can a site rank on Google and still be invisible to AI engines?
Yes, and it is common. Googlebot renders JavaScript and is whitelisted by default almost everywhere; AI crawlers mostly do neither. A page can sit on page one of Google while returning a blank shell or a 403 to GPTBot.
Does a high readiness score guarantee citations?
No. Readiness removes the blockers; citation also depends on grounding - whether engines trust and select your page over alternatives. The free audit measures that outcome directly, prompt by prompt.
What is the fastest readiness win?
Serve real HTML. Server-render or prerender your key pages so the answer exists in the initial response, then confirm retrieval bots are allowed through both robots.txt and your CDN.
Is llms.txt a ranking factor?
No one documents it as one. Treat it as agent guidance - a root file pointing AI systems at your canonical, high-value pages. Adoption is still early, and it costs minutes with no downside.
How is this different from the crawler policy check?
The crawler policy check reads sitewide robots.txt policy. This check scores a single URL across access, extraction, and guidance - the page-level half of the same question.
Explore the other free checks: