One crawl of your pages, read twice. Half of these checks are what any search engine needs before it can index and rank a page; half are what an answer engine needs before it can quote one. Here is every one of them, what it looks at, why it decides anything, and which the agent can fix for you.
Everything below comes from a single crawl of your pages. No connection, no API key, no snippet, no model — which is why it is free and why it runs before you install anything.
Half of it is what any search engine needs before it can index and rank a page. Half is what an answer engine needs before it can quote one. A few checks earn their place in both halves; those are marked. A check lives in exactly one section, so no number is ever counted twice.
An SEO retainer reports the first half, monthly. Almost nobody reports the second half at all.
Fourteen checks. If these fail, the page is not competing — it is not in the race.
A page that cannot be indexed cannot rank, however good it is.
noindex on a page removes it from Google entirely — it is the single most expensive one-line mistake on a website.What exists in the HTML before JavaScript runs, and how fast it arrives.
Structured data is what turns a page into facts a machine can use.
@id describe one company, not several unrelated fragments.The entity behind the site: one name, one company, linked consistently.
sameAs is how a machine confirms that the LinkedIn page and the website are the same organisation. Counts for both halves.Nine checks. These are the ones that decide whether an engine can reach your page, lift a passage from it cleanly, and find you in the sources it reads about your category.
The AI crawlers are separate from Google's, and are blocked separately.
robots.txt rules per AI fetcher — GPTBot, ClaudeBot, PerplexityBot and the rest. Allowing Googlebot says nothing about these.Answers are assembled from passages that stand on their own.
Whether the pages AI reads about your category mention you at all.
Each one comes back with what is wrong, the raw evidence behind it, and the
change that fixes it. Four of them the agent can apply itself with one click
from you — crawler access, Organization data, the sitemap, /llms.txt.
For several of the rest it writes the exact text and publishes it once you
approve. The remainder are handed to you as a specific instruction, because
some things genuinely need a person. What the agent can and cannot do →
Whether comparison articles mention you, and what the press and reviews about you say, are not things a crawler can settle. We mark those as judged by a person rather than guessing and presenting the guess as a measurement.
Yes, and it stays free because it costs us nothing meaningful to run: one crawl of your public pages and 23 deterministic checks. No model is called, no engine is asked anything, and a free account is all it takes to start.
Because every one of them changes a decision. We would rather run 23 checks whose failure you can act on than 120 that produce a score.
No. A number that blends a blocked crawler with a slow page tells you nothing about either. You get what failed, in the order that matters, with the evidence.
It says so, in one line, instead of passing or failing. Two checks — whether comparison articles mention you and what press and reviews say — are judged by a person rather than guessed.