About our checker
You are here because this user agent appeared in your logs: AgentsReadable-Check/1.0 (+https://agentsreadable.com/bot). It belongs to the instant Agent Readiness Check by AgentsReadable.
What it does
It reads public pages and files the way an AI agent would, scores how well agents can find, read and act on the business, and compares the site's facts with what an AI engine says about it. It never logs in, submits forms or changes anything, sends no cookies, and makes at most 15 requests per check.
What it reads
- The homepage (once, with our user agent — never retried with another user agent; once more over https:// if the site answered over http://)
- /robots.txt and the sitemap it declares, or /sitemap.xml
- The homepage with Accept: text/markdown
- /llms.txt
- /.well-known/api-catalog
- /index.md (only if no Markdown version was found)
- One pricing page linked from the homepage (only if the homepage shows no price)
- The homepage with the user agents of OAI-SearchBot, ClaudeBot, PerplexityBot (to detect CDN blocking)
- /.well-known/ucp
- /.well-known/mcp
- /.well-known/agent-card.json
- /.well-known/agent.json
How often
The checker is user-triggered: it fetches a site only when a person asks for a check of it, at most once per domain every 24 hours (repeat requests get the cached result). There is no scheduled crawling. Because it acts on a person's request, a wildcard User-agent: * rule in robots.txt does not stop it — the check reports such rules instead, so that site owners can see how agents are affected.
How to opt out
Add an explicit group for our user agent to robots.txt. The checker reads it right after the homepage and stops as soon as it sees the rule: the homepage response is discarded, nothing is scored and no result is stored.
User-agent: AgentsReadable-Check
Disallow: /Contact
Questions or removal requests: hello@agentsreadable.com. The method is published at /method; our own facts are at /facts.json.