Audits the four things a GEO owner actually controls on a live property: whether llms.txt exists and parses, whether robots.txt says anything to AI crawlers by name, whether the homepage publishes an entity graph, and whether discovery paths are declared. Every number below is fetched live when you press the button.
Score out of 100 across four controllable surfaces: llms.txt (25), AI-crawler directives (25), entity graph (25), discovery (25). Press the button to populate.
I tested this score against citation share I measured separately (82-prompt panel, Google AI Overviews, n=75). Spearman rho = +0.10 across the five scorable exchanges: no predictive relationship. OKX scores highest here and takes 5% of citations. Kraken leads citations at 25% and scores mid-table. Binance cannot be scored at all (its edge serves an empty 202) and leads citation share at 35%.
Use it to find configuration gaps (a missing file, an undeclared crawler, an absent entity graph), which it does accurately and reproducibly. Do not use it to rank competitors or forecast performance. A GEO tool that reports a confident composite score without stating this is selling you a number it has never validated.
robots.txt groups are most-specific-wins and they do not inherit. A crawler matched by its own User-agent group ignores every rule in *. A crawler with no group of its own sees only *. So crawl-hygiene rules written for Googlebot are invisible to GPTBot unless GPTBot is named too. This table counts the disallow rules each site wrote for named classic crawlers that no AI crawler will ever read.
| Domain | Classic crawlers named | AI crawlers named | Disallow rules AI bots never see | Rules in * |
|---|---|---|---|---|
| Run an audit to populate. | ||||
A high count in column four means the site invested in crawl hygiene and then exempted every AI crawler from it. Those crawlers spend their budget on the paths the site already decided were worthless.