Site and publishing
8
checks, all deterministic, none of them needing a model to decide
CheckID constants in internal/readiness — crawlers, llms.txt, Organization schema, FAQ schema, head basics, server rendering, about page, pricing page
This is not a site audit. Title lengths, alt text, and readability scores are Ahrefs' job and largely a commodity by now. The two gaps that actually cost citations, no directory profiles and no public pricing, are off-page findings and they live on Citation Sources instead.
What is left is the floor, and only the floor: eight things that either let an answer engine read you or stop it. Each one runs over your homepage and a handful of key pages, costs a few HTTP requests, and produces a verdict you can trace back to the bytes we fetched. We fail one of them ourselves.
Where it lives: Site audit in the dashboard, plus the free checker on /tools.
6
of 8 checks pass
brightmetric.com, read live. Every check names the file it read.
GPTBot, ClaudeBot, PerplexityBot, and Google-Extended are separate agents with separate purposes, and the one people block is usually not the one they meant to. The check reads your robots.txt and reports blanket rules you did not write, which is how most accidental blocks happen. The crawler checker runs that one check on its own, and the robots.txt generator writes the file if you need to start over.
llms.txt, Organization and FAQ JSON-LD, canonical and meta tags, and whether your copy is in the server response or arrives later via JavaScript. Also whether an about page and a pricing page exist, because those two are what an assistant reaches for when a buyer asks who you are and what you cost. Each of those has a generator or validator of its own: llms.txt, schema markup, and a structured data validator for what you already publish.
Pass, warn, fail, and unknown. Warn means nothing is broken and the engines are being given less than they could use, and it is never dressed up as a failure. Unknown means we could not fetch the page, which is a different fact from a broken site.
Included, and the eight-check version is free with no signup.
Read the full plan comparison.
Live demo
See Readiness checks on a real account, with no signup and nothing to configure.
Open the demoTraining crawlers and answer-engine crawlers are different agents. Blocking GPTBot does not remove you from ChatGPT's answers, and blocking the retrieval agents does. Most sites that meant to opt out of training have blocked the wrong one.
No engine has committed to reading it. It costs an hour, it is a plain markdown file, and it is the cheapest item on this list, which is roughly the honest case for it.
Same eight checks. The paid version stores runs against your product so you can see what changed, and sits next to the citation data that tells you whether being readable made any difference.
Free tools
The hub: what each free tool answers, and where the free version stops being enough.
Citation corpusCitation Sources
Every URL the engines cited in your category, classified, checked for whether you appear in it, and sorted by what you can actually win.
Site and publishingTo-dos
A focused task queue with evidence, practical steps, priority rationale and completion controls.
Start with the free visibility check to see whether the engines name you today, then run the scan that shows which pages they used instead.