AI traffic and crawler logs
Count visits that arrive from AI products without identifying anyone, and read which crawlers reach your server from an access log.
AI Visibility → AI traffic has two parts. Both are free and use only our servers.
Visits from AI products
A small script on your pages notices when a visitor arrives from an AI product (ChatGPT, Gemini,
Perplexity, Claude, Copilot, Kimi, Doubao, Yuanbao, Qwen, DeepSeek and a few others), from the
referrer or a utm_source. It sends the source name, the landing path (no query string) and
nothing else. The report shows visits by source and by landing page, beside your page changes and
the answers that cited your site.
- No cookie, no storage, no address, no browser details. Do Not Track and Global Privacy Control stop the script from sending anything.
- Not recorded: visits from the EU, EEA, UK and Switzerland, and visits whose place cannot be told. We decide on the visitor, not on where your company is, from the country our network edge reports (never from anything the browser says about itself); a European time zone in the browser can only make a visit not count. The feature is off for the whole platform until a legal review allows it, and off for a project until its owner turns it on.
- A floor. Many AI apps send no referrer at all, so the total is a lower bound. The report says so. Groups under 5 visits are merged into (other).
- Forgeable. The key is tied to your site's origin, which stops accidents but not someone who wants to send fake visits. Read the numbers as a guide.
- A day keeps at most 300 distinct landing paths; the rest are counted under (other). Data is kept 400 days. You can erase it at any time, which also replaces the key.
The report puts visits next to page changes and citations to show what happened at the same time. It does not say that one caused the other; for cause, compare pages you changed with pages you did not.
Crawler logs
Upload an access log (combined format, plain text or .gz, up to 20 MB; 100 MB and 200,000 lines
are read). You see which AI and search crawlers reach you, which paths they fetch and which are
turned away (401, 403, 429), for example by a CDN rule.
A crawler name in a log is only a claim. For Googlebot, Bingbot, Applebot and Baiduspider the address is checked by reverse DNS and a forward lookup back to the same address, and the result is shown as verified; for the rest, and when more than 200 addresses need checking, the entry stays a claim. Addresses are not stored, and neither is the file: only a summary per day, crawler, path and status. Five uploads a day.