About the scanner bot

User-Agent

Our scanner identifies itself as:

CanAISeeBot/1.0 (+https://canaisee.com/about-bot)

The scanner always identifies itself this way. It never sends other vendors' user-agent strings (GPTBot, ChatGPT-User, Claude-User) to test how a site treats them: on most edges that would be scored as impersonation, and it would disqualify the scanner from verified-bot programs.

What it fetches

When a visitor enters your URL, the scanner fetches a bounded set of resources from your site in parallel: the canonical URL (plain, and once more with Accept: text/markdown), /robots.txt, /sitemap.xml (plus up to 20 child sitemaps), /llms.txt, /llms-full.txt, the markdown alternate the page advertises, and the discovery files agents read: /.well-known/ai-catalog.json (and up to three cards it links on your own host), /.well-known/agent-card.json, /.well-known/mcp/server-card.json, /.well-known/agent-skills/index.json, /.well-known/api-catalog, /.well-known/oauth-protected-resource, /.well-known/ucp, and /.well-known/acp.json. For stores whose page carries no Product markup it samples one product page from the sitemap. For one check, it additionally loads the canonical URL in a headless Chromium instance to compare the rendered text to the plain HTML.

Rate and politeness

The scanner honors robots.txt. Each scan produces a bounded set of requests against your origin (typically 15–25, all GET), and a per-URL cooldown of five minutes prevents repeated scans from re-fetching.

Blocking the scanner

Add this to your robots.txt to decline:

User-agent: CanAISeeBot
Disallow: /

Takedown for a specific scan

If a scorecard permalink exists for your site and you want it removed from the public archive, email takedowns@canaisee.com with the URL.