The ai-discovery radar
We measure what machines actually find on websites — and publish it, including the uncomfortable parts. A great deal is claimed about AI visibility and very little is measured. Yet the routes a website offers machines, so that it can be found and read correctly, can simply be counted. That is what we do: monthly, in public, with the method laid open.
What gets measured
The radar checks which standardised files a website makes available to machines: robots.txt, which for thirty years has governed where automated visitors may go. llms.txt, a table of contents written for language models. ai-catalog.json, which describes the content and services a domain offers AI agents. 33 such routes are in the running measurement.
Then there is what we deliberately do not measure. Another 12 routes sit under observation: known to us, but not yet widespread enough to justify a request on every host in the sample. And 49 we examined and rejected, with a reason recorded entry by entry. A catalogue does not get better by admitting every format someone happens to mention.
Only public configuration files meant for machines are fetched — no page content, no images, no text. The measurement obeys each domain's robots.txt and identifies itself by name.
What the first run shows
As of August 2026, measured across a stratified sample of 853 reachable domains:
- 77.0% serve a
robots.txt. The oldest standard is the only one nearly everyone honours. - 10.2% serve an
llms.txt. The format is young, and adoption reflects that. - 0.0% serve an
ai-catalog.json. Of the 33 routes probed, 16 returned not a single response across the entire panel.
Those zeros are not a measurement error. They are the result. Many of these formats are discussed as though they were established. On the open web, they are not.
Why we publish this
We build websites meant to be found by people and by AI systems alike. That work needs numbers instead of assumptions: which route actually carries weight today, and which is a bet on tomorrow? Anyone who does not measure is selling guesswork.
So the method is public — sample, ruleset, confidence intervals, and the routes that came back at zero stay in the table. Anyone who doubts one of these figures can recompute it. That is the whole point.
Who does not want to be measured
No questions, no reason needed. An email with the domain is enough, or a single line in your own robots.txt. Excluded domains are skipped before any request is made.
Questions about the measurement?
Write to us — about the method, about a single figure, or if your domain should not be measured.
- Reply usually within 24 hours
- Opt out with no questions asked
- Method fully disclosed