AI is reading this site.
Here’s the live tape.
Every AI bot fetch on garnetgrid.com — OpenAI, Anthropic, Perplexity, Google, Apple, Amazon, ByteDance — counted at the edge and published daily. We productize citation engineering for clients. We measure it on ourselves first.
Top vendors — last 30 days
Top bots — last 30 days
Most-fetched pages — last 30 days
What you’re looking at. Every request to a garnetgrid.com page passes through a Cloudflare Pages middleware tap that checks the User-Agent against ~26 known AI bots (training crawlers and retrieval bots both). Matches get logged to Workers Analytics Engine. A daily cron rolls 24 hours of data into crawler-public.json — the same blob this page renders.
Why publish it. The four-engine panel (ChatGPT / Claude / Perplexity / Gemini) cites the sites it indexes. If you’re selling to anyone an AI assistant might recommend, you should care which engines fetch your pages. This is the same data we produce monthly for paying GEO subscribers.
The dataset is CC0. Pull /assets/data/crawler-public.json directly. No auth, no rate limit. Schema documented at geo-methodology.