Access & crawlability
30 pts
- AI bots allowed in robots.txtproven10
- Real fetches as AI bots succeedproven10
- Sitemap present and truthfulproven5
- Missing pages return real 404shygiene5
AI Readiness Check · free, ungated
Your buyers ask ChatGPT. Their agents try to book you. 15 live checks tell you whether the machines can read you at all, with evidence labels no other checker is honest enough to print.
No email · no signup · scored in seconds. Scan a site you own or are evaluating.
The rubric
100 points across four categories, weighted by evidence, not by what other checkers happen to score. Every check is labeled: proven affects what AI engines can do today; hygiene is correct and cheap; speculative scores points on every tool in the category (including this one) without citation evidence behind it. Competitors can’t copy that last label without indicting their own scoring.
Access & crawlability
30 pts
Content & rendering
30 pts
Structured data & entity
25 pts
Agent usability
15 pts
Why the labels exist
Every checker in this category scores llms.txt like it moves citations. The evidence says it doesn’t; no major answer engine consumes it today. We still check it (3 points, labeled speculative, cheap to fix) — but a tool that won’t tell you which of its own points are theater isn’t measuring your site. It’s measuring your anxiety.
Same rule for the score itself: 95+ means mechanically AI-ready, not “cited by ChatGPT.” Readiness is the entry ticket; citations take authority and content over time. We’d rather you hear that from the tool than from a sales call.
FAQ
15 live checks across four categories: access & crawlability (robots.txt AI-bot rules, real fetches as GPTBot and PerplexityBot, sitemap health, soft-404s), content & rendering (server-rendered text, content depth, token weight, first-byte speed), structured data & entity (JSON-LD validity, organization identity, sameAs entity links, meta/OG), and agent usability (an agent-actionable booking path, llms.txt, markdown availability). The scan fetches everything live from your site rather than reading from a cache of old crawls.
Free and ungated: no email, no signup. The catch is honest: we build AI-ready websites for a living, and a tool that shows you exactly what's broken is the best introduction to how we work. The score is yours either way.
Every check carries an evidence label. Proven: it affects what AI engines can actually do today; a blocked bot can't read you, and JS-only content is invisible to most crawlers. Hygiene: correct and cheap with a plausible mechanism, worth doing. Speculative: it scores points on every checker (including this one) without citation evidence behind it; llms.txt is the famous example, and we say so. Other tools weight all of these as if they were equal. They aren't.
No, and any tool that implies it does is overclaiming. A 95+ score means your site is mechanically AI-ready: engines can crawl it, read it, understand who you are, and agents can act on it. Being cited and recommended also takes authority and content built over time. But without readiness, none of the rest can start.
The robots.txt check covers 18 tokens across every major engine: OpenAI (GPTBot, OAI-SearchBot, ChatGPT-User), Anthropic (ClaudeBot, Claude-SearchBot, Claude-User), Perplexity (PerplexityBot, Perplexity-User), Google-Extended, Bingbot, Meta (Meta-ExternalAgent, Meta-ExternalFetcher), MistralAI-User, Amazonbot, Applebot, Applebot-Extended, DuckAssistBot, and CCBot. Worth knowing: Google-Extended governs Gemini training and grounding only; AI Overviews ride on plain Googlebot. On top of the parsing, we make real fetches as GPTBot and PerplexityBot and compare them to a browser fetch, which catches the WAF and edge-level blocking that robots.txt parsing can never see.
The result is stored for 90 days so your score link keeps working, and repeat scans of the same URL are served from a 1-hour cache so the tool can't be used to hammer anyone's site. We don't sell your data. The optional email field sends you the full report once you confirm the address. Updates from us are a separate, unticked box: you only get them if you tick it, and every one has a one-click unsubscribe. Details are in our privacy policy.
Every failed check comes with a copy-paste fix snippet, matched to your stack where we can detect it (Next.js, WordPress, Shopify). If you'd rather not do it yourself: we build sites that ship scoring 95+ on this exact check, on fixed tiers at $3K / $7.5K / from $15K, live from 2 weeks. The rubric is public, so you can hold us to it.
From the team behind the rubric
AI-ready marketing and brand sites at $3K, $7.5K, and from $15K, live from 2 weeks, engineered for the exact layer this check measures. Every build runs the check at handover, and the rubric stays public.