Our own review, published so it can be checked

We reviewed 29 free AI crawler checkers. 11 treat GPTBot as ChatGPT's search crawler.

An AI crawler checker tells a website owner which AI crawlers can read their site. We reviewed 29 free ones and compared what each says about three crawlers with what the crawlers' makers say. 11 treat GPTBot, OpenAI's training crawler, as the crawler behind ChatGPT search. OpenAI says that job belongs to OAI-SearchBot.

By Vector Index AI. Review carried out 16 October 2026. Published 20 October 2026.

Three things the checkers get wrong

Each row counts the checkers that said it, out of 29. A checker counts once per row however many times it says it.

What the checker says	What the maker says	Checkers
GPTBot is ChatGPT's search or citation crawler	OpenAI says GPTBot "is used to crawl content that may be used in training our generative AI foundation models", and OAI-SearchBot "is used to surface websites in search results in ChatGPT's search features" (OpenAI).	11
Google-Extended controls AI Overviews	Google says Google-Extended "does not impact a site's inclusion in Google Search nor is it used as a ranking signal" (Google). AI Overviews are part of Google Search.	6
Fetchers that visit a page because a person asked, such as ChatGPT-User, follow robots.txt	OpenAI says of ChatGPT-User: "Because these actions are initiated by a user, robots.txt rules may not apply." Perplexity says Perplexity-User "generally ignores robots.txt rules" (Perplexity).	3

The counts are strict. Wording that comes close without saying it outright is left out: 2 more checkers on GPTBot, 2 more on Google-Extended, and 7 more that imply user fetchers follow robots.txt without saying so. Each count is a floor.

Why it matters to a site owner: a checker that reads GPTBot as ChatGPT search can tell someone who blocked training, and nothing else, that they are invisible in ChatGPT search. A checker that reads Google-Extended as the AI Overviews switch points them at the wrong line.

What else we saw in the 12 we ran

Reading the file. Of the 11 checkers we could run, 10 read our test robots.txt files correctly. One reported 10 of its 27 crawlers as blocked on our own site, where the robots.txt standard (RFC 9309) allows all 10 at the home page. It marks a crawler blocked when any Disallow line appears anywhere after the crawler's name.
A fix that opens more than it fixes. One checker's generated "recommended robots.txt" moved our four private-path Disallow lines into the * section and gave each named crawler only Allow: /. A crawler with its own section ignores the * section, so pasting it would have opened those four paths to all 18 crawlers it names.
Server logs. 2 of the 12 read server logs. Neither checked a visit's IP address against the addresses the crawler's maker publishes. We gave both a test log with crawler names on addresses outside those lists, and neither flagged any of them.

How we reviewed them

17 checkers, from their pages. Each tool page was fetched and its text and FAQ answers read. We did not run these 17, so what we report for them is what their pages say they do.
12 checkers, run. Each page was loaded in a browser and, where the tool allowed it without signing up, run on vectorindexai.com, our own site. Tools that only take pasted text got test files. 11 of the 12 ran. One blocked our automated browser, so for it we read the page only.
Against the makers' own pages. What each checker says about GPTBot, Google-Extended and user fetchers was compared with OpenAI's, Google's and Perplexity's own crawler documentation.
Nothing entered. No email address was given to any tool, and no paid tier was used.

The 29 checkers

Reviewed from their pages: Jahanzaib.ai (jahanzaib.ai), Lumina SEO (lumina-seo.com), Tufan Erdogan (tufanerdogan.com), AI-Advisors (ai-advisors.ai), AEOVisor (aeovisor.com), AEO.page (aeo.page), Citedit (citedit.com), Kitbase (kitbase.dev), Decoding (crawler.trydecoding.com), CheckAIBots (checkaibots.com), PromptDNS (promptdns.com), Backlinko / Semrush (backlinko.com), 365i (365i.co.uk), GEO Toolbox (geotoolbox.ai), LLM Pulse (llmpulse.ai), LLMrefs (llmrefs.com) and AIclicks (aiclicks.io).

Loaded and run: MRS Digital (mrs.digital), SEOCrawl (seocrawl.ai), isvisible.ai, xSeek (xseek.io), Prefer (tryprefer.com), AI Crawler Check by Horatos.ai (aicrawlercheck.com), Zeevron (zeevron.com), Beamtrace (beamtrace.com), RankX AI (rankxai.com, page only), AI Bot Access (aibotaccess.com), Rank Prompt (rankprompt.com) and Growtika (growtika.com).

This page reports totals. It does not say which checker made which error.

Limits

One day. Reviewed 16 October 2026. Checkers change, and any of them may have been corrected since.
Pages, not runs, for 17. For the first 17, a feature or claim is what the page says, not what we saw the tool do.
One reader. Each call is one reviewer's reading of the page against the makers' wording. Borderline cases are left out, so the counts are floors.
Not tested. How any checker handles wildcard rules (* and $), any paid tier, and one checker's live results, because it blocked automation.
We are not neutral. Vector Index AI makes a free AI crawler checker. It labels GPTBot as a training crawler, says Google-Extended doesn't affect Google Search, and says robots.txt may not apply to user fetchers.