Can a machine read this?
A free scanner: paste a URL, and 30 seconds later you get one score out of ten for how readable your site is to an AI agent — with the specific misses named.
- built with
- Firecrawl
- Claude Code
- Railway
- robots + schema parsing
- free, no signup
Paste a URL. One number back in 30 seconds. No signup.
Machines read your site. Yours is invisible.
whyclose
The new readers want a clean DOM, llms.txt, schema, semantic URLs. Not your animations. Nobody was checking, so nobody knew.
Thirty seconds, start to answer.
urlYou paste an address.
No account, no email, no sales call. That was the point — anything with a signup would not get used.
robotsIt reads what you allow in.
robots.txt and the meta directives first, because a crawler that is blocked never gets to the rest.
structureThen how the page is built.
Semantic markup, schema.org, headings that mean something, llms.txt if you have one.
scoreOne number. Most land under 5.
Out of ten, with the specific misses named. A number people argue with is a number people act on.
A week later, they shipped ours.
whyclose
We were already out, with a report on large Swedish names failing the basics — and scoped to Nordic and EU rules, which is the part the American tools leave out.
We killed the API scanner. It lied.
whyclose
Too many ifs and buts: it returned the wrong answer on real sites. We shipped less rather than ship something wrong.
Small tool. Real use.
The whole thing, in four lines.
- the gapTwenty years of optimising for Google, and nobody had checked what an AI agent actually sees when it reads a site.
- the buildPaste a URL, one number back in 30 seconds. Scoped to Nordic and EU rules, which the American tools skip.
- the raceA week after we shipped, a far better-funded team launched the same idea. We were already out with a report.
- the cutWe killed the API scanner because it returned wrong answers on real sites.