GETCITEDMeasured, not generated.
Every number below comes from a crawl, a search result or an engine's answer. Here is what we look at, and what each figure means.
We fetch your site the way an answer engine does.
No JavaScript executed, robots.txt respected, same host only. Each rule reports the measurement it fired on.
We ask the engines your questions and record who they cite.
Your tracked prompts run on 5 engines. For each answer we keep the citation, the brand mention and the domains named instead of you.
Position over time, and whether the AI Overview cited you.
The sub-queries engines issue for your prompts become tracked keywords of their own, because ranking for those is what gets you retrieved.
Where visitors stall, so you can adjust your funnel.
A cookieless snippet, aggregates only, and a difference is only called one when the sample is big enough to support it.
A problem, its evidence, and no opinion about your code.
No proposed titles, no rewritten copy, no patches. Your agent knows your codebase and your product; we do not.
robots.txt disallows GPTBot, ClaudeBot and PerplexityBot. Nothing you publish can be retrieved by them.
One key, and it works the queue from your editor.
Run the audit, read a finding with its evidence, check rankings, check AI visibility, report back what changed.
Or none of that, and just use the dashboard.
Run the audit on a button, filter the queue, read any finding. Fix something in your editor and the next crawl closes it for you: there is no “mark as done” button, because that would record an opinion where a crawl records a measurement.
"Auth page, deliberately noindexed. Adding words here would help nobody."
your agent · example