Can AI actually read your site?
See whether GPTBot, ClaudeBot, PerplexityBot and the rest are allowed to read your site.
Free, no signup. Reads robots.txt and requests your homepage as each crawler.
If a crawler is blocked
Access is the floor, so confirm it is really blocked before you change anything.
How the biggest sites answer the same question
We ran this check against the most visited sites on the web and published every verdict with the robots.txt line responsible. The per-domain index is worth a look before you change anything: plenty of very large sites block a training crawler on purpose and leave the one that decides citation alone, and seeing which is which is faster than reasoning about it.
Common questions
Ten agents across five engines. Each is asked for your homepage by name, and its robots.txt rules are read separately, because a site can permit a crawler in one and refuse it in the other. ChatGPT: GPTBot trains and grounds OpenAI models; OAI-SearchBot builds ChatGPT's search index; ChatGPT-User fetches a page when a user's question needs it. Claude: ClaudeBot trains and grounds Anthropic models; Claude-User fetches a page during a Claude conversation; Claude-SearchBot builds Claude's search index. Perplexity: PerplexityBot builds Perplexity's index; Perplexity-User fetches a page to answer a live question. Gemini: Google-Extended controls Gemini grounding and training; Googlebot feeds Google's index, including AI Overviews. Google's two agents gate both Gemini and AI Overviews, so they are a single edit in robots.txt covering two surfaces.
GPTBot is OpenAI's crawler. It reads pages to train and ground OpenAI's models, including ChatGPT. Blocking it in robots.txt removes your site from that pool.
Blocking GPTBot, ClaudeBot or PerplexityBot has no effect on Google Search, because those are different crawlers. Google-Extended is the exception worth understanding: it controls Gemini grounding and training but does not affect Google Search ranking either. Googlebot itself is the one that does both, so disallowing it removes you from Search and from AI Overviews together.
robots.txt is a request, not a wall. A CDN or firewall rule that rejects anything without a browser fingerprint will return 403 to a crawler regardless of what robots.txt says. That is why this tool also requests your homepage as each agent instead of only reading the file.
No. Access is necessary, not sufficient. It means a crawler is permitted to read you; whether an engine then names you in an answer depends on what it finds and what else it has read.
What changes if you track it
A tracked domain re-runs this check every week and tells you when a rule changes, which is the case that matters: access is usually fine on the day you look and breaks on a deploy nobody connected to it. See plans.
More free tools
Access is the floor. The check is the measurement.
This tool answers whether an engine is allowed to read you. The free check answers what those engines say when a buyer asks in your category, scored with the band it is accurate to.