Free tool
Half of SEO is about being found. The other half — the half that decides whether an AI system can use what it found — is what this audits. Enter a URL and you get a scored report over 25 checks, with the raw evidence behind every finding.
Anonymous checks are limited per day. We respect robots.txt and identify
ourselves in the User-Agent.
25 checks in five areas. The weights are fixed and published, so a score means the same thing on every site.
| Area | Weight | What it covers |
|---|---|---|
| Crawlability (8) | 25% | robots.txt rules per crawler, sitemap, redirects, status codes — whether a crawler can get in at all |
| Basics (7) | 25% | title, meta description, canonical, headings, image alt text |
| Content readability (4) | 25% | how much text exists without JavaScript, and how much sits behind it |
| Structured data (4) | 15% | JSON-LD entity declarations, language attributes |
| Other technical checks (2) | 10% | the slow and roundabout responses that reduce how often you get revisited |
Every check is documented in public: what it looks for, why it matters, how to fix it. The engine is not open source; the rules are.
Not one number for the whole site — a site can pass on the homepage and fail everywhere else.
Each with the evidence: status codes, byte counts, the exact HTML
fragment. Every conclusion can be reproduced with curl.
A corrected robots.txt, an llms.txt,
JSON-LD, canonical tags — plus a written brief you can hand to an AI assistant.
Printable to PDF, in plain language. If you are the one delivering the audit rather than receiving it, this is the part you send.
Saying this plainly is cheaper than having you find out later: there are no backlinks, no keyword rankings, no Core Web Vitals, no content quality judgement, and no prediction of whether a specific AI will cite you. The engine reads your pages the way a crawler does — it does not run your site in a browser or measure how fast it feels.
That narrowing is the point. Most "AI SEO" advice tells you to publish more; none of it helps if the crawler cannot read what you already have. This audits the precondition.
The same check, framed around which crawlers can read you, with the full user-agent list.
Every token worth knowing, and what blocking each one costs.
Which of the files AI reads does what.