Cita is a set of free skills and a connector for Claude, a command for the terminal, and this website as the record. Start with the free check: it says whether AI assistants can read your site at all, and it costs nothing.
Measuring endurance & outdoor apparel today — more categories as they earn their instrument.
Correlational monitoring signal · not a causal promise
Run the free check, open the board and the source map. Nothing to install, nothing to sign up for.
Add the Cita connector and the free skills, then ask in plain words: “can AI read my site?”, “what does ChatGPT read for trail running shoes?”, “what changed since last time?”
The same features as commands: cita check, cita map, cita board, cita sources. No key, nothing spent.
Six things, by the same names everywhere. None of them is a score.
Can AI read your site? Checks whether the crawlers behind ChatGPT, Claude, Perplexity and Google's AI answers are let in, and whether a sitemap and structured data are there.
It does not say that being readable gets you cited; it rules the plumbing in or out, nothing more.
Map the sources ChatGPT cites in your category, stage by stage, from the buyer's first question to the last.
It says nothing about one brand's position; it is brand-free by construction.
The publications ChatGPT cited most this week, per category, re-measured every Monday.
It is not a ranking of quality or influence; it counts citations in the answers we read.
See where ChatGPT cites your competitors, not you: ask an AI assistant your buyers' questions, several times, and keep who it named and what it read.
It is not a score, and one reading is not a trend; one reading is a dated snapshot. A reading taken with Claude's own web search says it is Claude-only.
What entered and left the assistant's reading list since last time, with the noise floor beside every arrow.
It does not say a change was caused by anything you did; movement below the floor is noise.
Fetch the cited pages (up to 40 per audit, newest first) and check each one, so every gap is dated and checkable.
It does not say that fixing a page will get you cited, and a page it could not fetch is reported as not checked, never as broken.
What was not measured is reported as unmeasured, never as zero.
The source map shows which publications ChatGPT cites for the endurance categories, stage by stage. The live board has no measurement to show right now, so it is not offered here as proof.