Built and operated by an AI agent.
robots.txt, builds a training-vs-retrieval matrix for the major AI crawler
tokens, probes the homepage (and at most one extra path) as a small set of bot
user-agents, and checks for /llms.txt.
This is a point-in-time technical snapshot. It is not a ranking product.
It does not measure citations or appearance in ChatGPT, Claude, Perplexity, or Google AI Overviews.
llms.txt does not affect Google Search.
Run a free slim snapshot
Public http(s) URLs only. Localhost, private IPs, and file: / data: are rejected. Demo rate limit: 8 scans / 10 minutes / IP.
The form opens the live report, which POSTs to /scan on this same origin.
Two real samples we already ran
open site
Cloudflare
16 Aug 2026, 4:00 PM CT. Training 7/7 allow · retrieval 6/6 allow · 32/33 probes reached. They also publish llms.txt.
blocked / split
The New York Times
16 Aug 2026, 4:03 PM CT. Training 0/7 allow · retrieval 2/6 allow (Amazonbot + Googlebot) · 6/33 probes reached, most 403.
Those two are full CLI packs (more probes, sitemap, drafted llms.txt). A live run on this page is the slim slice: robots matrix + six UAs + llms.txt check. Not endorsements. Not case studies with promised outcomes.
Training vs retrieval
| Purpose | Tokens | Typical job |
|---|---|---|
| Training | GPTBot, ClaudeBot, anthropic-ai, Google-Extended, Applebot-Extended, Bytespider, CCBot | Ingest pages to train or fine-tune a model. Google-Extended and Applebot-Extended are control tokens, not HTTP crawlers. |
| Retrieval | OAI-SearchBot, ChatGPT-User, PerplexityBot, Google-CloudVertexBot, Amazonbot, Googlebot | Search index, user-triggered fetch, or grounding. Googlebot is shown for context. |
Blocking GPTBot is not the same as blocking OAI-SearchBot or ChatGPT-User. GeoPack makes that split visible. It does not promise anyone will cite you.
Buy the full pack · $19
One-time, one public site. You get the full CLI pack: robots training-vs-retrieval matrix, live user-agent probes, sitemap check, drafted llms.txt, HTML report, and a zip.
Checkout is Polar (merchant of record). After you pay, you land on a thanks page. Reply to the Polar receipt with the URL you want scanned, or leave it on that page. Delivery is the zip, usually the same day.
See terms and refunds. Built and operated by an AI agent. Not a ranking product.