Can ChatGPT, Claude and Perplexity read your website?
AI assistants fetch pages with their own bots before they answer or cite anyone. An old robots.txt line, a page that only renders with JavaScript or a missing llms.txt can keep you out of their answers. Check in ten seconds, free.
What the score means
100 points over 10 checks. The biggest share (25) is whether the bots that answer people may read your site: OAI-SearchBot and ChatGPT-User, Claude-SearchBot and Claude-User, PerplexityBot, Googlebot, Bingbot and others. Blocking training bots (GPTBot, ClaudeBot, Google-Extended) does not cost you points; that's a fair choice and separate from showing up in answers.
The rest covers whether the home page loads for a plain client (15), has text without JavaScript (10), and offers llms.txt (15), a sitemap (10), robots.txt (5), a title and description (5), schema.org data (5), no noindex/noai tag (5) and agent endpoints such as an MCP server card or OpenAPI file (5). A is 85 and up, B 70, C 55, D 40.
Check many sites, or from your agent
The same check is an API and MCP tool for $0.005 a site, with 20 free calls a day on a free key. Agencies use it to audit client lists; agents use it before they rely on a site.
curl -X POST https://api.exepal.com/v1/agent-ready \
-H "Content-Type: application/json" -H "X-Exepal-Trial: 1" \
-d '{"domain":"acme.com"}'
API docs, MCP setup and output fields
Questions
- Is this the same as SEO?
- It overlaps, but it's narrower: can AI bots get in and read the page at all. Whether an assistant then recommends you is a separate question, which the free AI visibility check answers.
- Why does my site say "not reachable" when it works in my browser?
- Bot protection or a firewall is likely blocking non-browser clients. AI assistants hit the same wall, so it's worth allowing the verified AI search and assistant bots.
- How fresh is the result?
- Results are kept for 6 hours so share links and badges load fast. Fix something and check again after that to see the new score.
- Do you store my site's data?
- Only the result, for 6 hours. We read the same public files any crawler reads: robots.txt, sitemap, llms.txt, the home page and a few /.well-known files.