AI Visibility Checker

Ask AI search the questions your customers ask — see whether it cites your site, which sites it recommends instead, and whether robots.txt keeps AI bots out.

Questions a customer would ask an AI (up to 3)
1
2
3
AI answer checks today: 0 of 10. Repeating the same questions within 24 hours does not use the limit.

AI visibility on your real queries — SEO Radar

Here you make up the questions. Radar takes the queries your site already gets Google impressions for (from your Search Console), asks them to AI search and shows whether it cites you and whom it cites instead; every check is kept in a history. With Google Analytics 4 connected, Radar shows visits from ChatGPT, Perplexity, Gemini and Claude and which pages they land on — and finds pages just below Google’s top.

Open SEO Radar

Free AI Search Visibility Checker

This tool shows whether AI search can see your site. Enter up to three questions a customer would ask — or let us draft them from your homepage — and we put them to Perplexity Sonar, an AI search engine with live web results and a list of sources. For each answer you see whether your site is cited as a source, whether your brand or domain is named in the text, how the answer begins and which sites the AI cites instead of you. The second part of the report is free and uses no AI: homepage access for 14 crawlers — GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, Claude-User, PerplexityBot, Perplexity-User, Google-Extended, Applebot-Extended, CCBot, Bytespider, meta-externalagent and Amazonbot — evaluated under the robots.txt standard, whether /llms.txt exists, noindex and nosnippet directives, and a test request with the GPTBot user agent that shows whether a firewall or CDN blocks AI bots. Answers to the same questions are kept for 24 hours, and repeating them does not use your limit.

Whether AI search cites your site when customers ask questions
Which sites the AI cites instead of you — per question and overall
Brand mentions in the answer text, even without a link
Homepage access for 14 AI crawlers in robots.txt: search, training and user-triggered fetches
Whether a firewall or Cloudflare blocks GPTBot; llms.txt, noindex, nosnippet
Free: question ideas drafted from your homepage

How to Use

  1. Enter your website and, optionally, your brand name
  2. Type up to three questions the way a customer would ask, or click “Suggest questions from my site”
  3. Run the check and see whom the AI cites and whether your site lets AI crawlers in

FAQ

No. The questions go to Perplexity Sonar, an AI search engine that searches the web at answer time and returns its sources. It is a good proxy for whether search-enabled AI assistants can see your site, but ChatGPT, Gemini and Copilot use their own indexes and may answer differently. Access for OpenAI, Anthropic and Google bots is checked separately, via robots.txt.

The AI builds each answer from scratch: the pages it finds, the wording and the sources all change. So one check is a snapshot, not a ranking. To see a trend, ask the same questions regularly and compare the share of answers that cite you. Here, repeating the same questions within 24 hours shows the saved answer.

According to OpenAI’s documentation, no: GPTBot collects training data, while OAI-SearchBot decides whether you appear in ChatGPT search. ClaudeBot and Claude-SearchBot are split the same way. If you want AI assistants to cite your site, keep the search bots allowed; whether to block training is your call.

No. Google-Extended is not a separate crawler but a robots.txt token: it stops the site from being used to train Gemini models and to ground answers in Gemini apps. It does not affect Google Search, including AI Overviews and AI Mode — those follow the regular Googlebot rules and meta tags such as nosnippet.

Without an account — 10 AI answer checks a day, with a free account — 30, with Pro — unlimited. Repeating the same questions within 24 hours does not use the limit. Each question is a paid AI request, so the tool also has a shared daily limit for the whole site; crawler access via robots.txt is always checked.

What AI search visibility means

When someone asks ChatGPT, Perplexity, Gemini or Copilot “which service should I pick” or “where can I buy”, a search-enabled assistant finds a handful of pages, summarises them and shows links to its sources. AI search visibility is how often your site ends up among those sources, or is at least named in the answer. There is no ranking position: every answer is assembled from scratch, so what you measure is the share of answers that cite you.

An AI answer has two layers. The first is what the model learned in training — it may know your brand without searching at all. The second is retrieval at answer time: the model sends queries to its own or a partner’s web index, reads the pages it finds and links to them. Links to your site almost always come from the second layer, and that is the layer this checker tests.

Which AI crawlers do what

The most common mistake is blocking “all AI bots” in robots.txt and disappearing from assistants’ answers as a side effect. The big operators split their bots by job:

JobBotsWhat happens if you block them
Assistant search indexOAI-SearchBot, Claude-SearchBot, PerplexityBotThe assistant cannot find or cite your pages
User-triggered fetchChatGPT-User, Claude-User, Perplexity-UserThe assistant will not open a link a user gave it (some operators do not apply robots.txt to these fetches)
Model trainingGPTBot, ClaudeBot, CCBot, Bytespider, meta-externalagent, AmazonbotFuture models will know less about your site; no direct effect on links in search
Control tokensGoogle-Extended, Applebot-ExtendedOpts out of Gemini training and Gemini answers, or Apple model training; Google Search and AI Overviews are not affected

A “don’t train on my site, but do cite it” setup looks like this:

User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

User-agent: OAI-SearchBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

Remember that a bot follows only the group with its own name: once GPTBot has its own group, the rules under “User-agent: *” no longer apply to it. Beyond robots.txt, your firewall can block bots too — Cloudflare has separate AI crawler blocking settings that work on top of robots.txt. That is why the checker also sends a test request with the GPTBot user agent and compares the response with a regular browser’s.

How to pick questions to test

The questions shape the result more than anything else. A few rules:

  • Leave out your brand. “What is [your company]” will almost always lead to your own site and tells you nothing about the competition. Test the questions of someone who has not heard of you yet.
  • Conversational, not keyword-style. Not “flower delivery London” but “what is the best flower delivery service in London and why”.
  • Mix intents. One “best…” question, one “how do I…” problem question, one comparison (“A or B — which is better”).
  • Your customers’ language and location. In English and in Polish the AI searches different parts of the web and cites different sites.
  • Real queries. The best source is Search Console: long, question-like queries your site already gets Google impressions for. Those are exactly what SEO Radar uses.

How to read the result

Your site is cited

The AI found your page and treated it as a reliable source. Check which page was cited: often it is not the one you are promoting but a blog post or FAQ. That tells you which format the AI finds useful.

Brand named, but not linked

The model knows you from reviews, directories or training, but used other pages to build the answer. Usually you are missing a page of your own that answers this exact question directly and briefly.

Not there at all

Rule out technical issues first: are AI search bots allowed, is there a noindex, does the site respond to bots? Then look at the sites the AI cites instead of you. If they are reviews and roundups, the way in is getting mentioned there. If they are direct competitors, compare their pages with yours: structure, completeness of the answer, freshness.

What helps you get cited

  • Regular SEO. AI search picks sources from a web index, and pages that rank at the top of classic search are cited noticeably more often. Indexing, rankings and links remain the foundation — check them with the Google rank checker.
  • Answer first. A short, direct definition or conclusion at the start of each section, details after. Models prefer passages that make sense out of context.
  • Specifics. Numbers, dates, prices, terms and comparison tables — things that can be quoted.
  • Mentions on other sites. “Best of” lists, reviews, industry directories and forums are frequent sources of AI answers. The report shows you exactly which domains those are.
  • Freshness. For “in 2026” questions the AI prefers updated pages with a visible date.

A detailed walkthrough with examples is in how to get into AI answers.

How to track AI visibility over time

A one-off check is a snapshot. For trends there are free sources: Search Console now has a generative AI features report with AI Overviews and AI Mode impressions, Bing Webmaster Tools reports citations in Copilot, and Google Analytics 4 has a separate channel for visits from AI assistants. Paid trackers ask the same set of questions every day across several AIs; prices and features are compared in our AI visibility tracking tools review.

If your site is connected to Search Console, SEO Radar asks AI search your real queries, keeps a history of checks and, with Google Analytics 4, shows visits from ChatGPT, Perplexity, Gemini and Claude by page.

Cascade link promotion for your site $30 bonusBacklinks 5→50→5000 across 100+ platforms. From $30 to start, plus a $30 bonus for newcomers.