Free AI visibility checker
When buyers ask AI about your topic, does it name you?
We ask Claude and GPT the questions your buyers ask, then show you whether they name your brand, where you come in the answer, and who they name instead.
What this AI visibility checker does
More buyers now start with a question to an AI assistant instead of a search box: “what are the best running shoes for beginners”, “which accounting software should a small agency use”. The answer names a handful of brands. If yours isn’t one of them, that buyer may never hear of you.
This free tool checks that for your brand. Give us your brand name, your website and your topic in a few words. We write six realistic buyer questions for that topic, the kinds people actually ask: the best options, a recommendation for their situation, the best value, how to choose, a problem to solve, and how the main options compare. None of the questions mention you, because buyers who haven’t heard of you won’t either.
We then put those questions to two AI models and read every answer for your name. You get a visibility score, the answers question by question, the brands named most often, and a short list of things to do about it.
How to read your result
- Visibility score
- The share of questions where at least one answer named you. Six questions, so each one is worth about 17 points.
- Position
- Where you came among the brands an answer named, in the order they appear. #1 of 7 means you were named first, out of seven.
- Named instead
- The brands an answer named when it didn't name you. These are the names a buyer hears in your place.
- Cited your site
- A web-search answer linked to your domain as a source. Useful, but it isn't the same as being recommended by name.
A low score doesn’t mean your product is worse. It means the models haven’t connected your name with these questions yet. A high score with a late position means you’re known, but other brands come to mind first. The brands named instead are often the most useful part: they tell you whose roundups, comparisons and reviews you need to be part of.
How it’s worked out
Every question goes to Claude Sonnet 5, answering from what it learned in training. The first two also go to GPT-5.6 Luna with web search, which looks at current pages before it answers. So you see both sides: what a model remembers about your topic, and what it finds when it searches today.
Whether you were named is decided by code, not by asking a model to judge. We match your brand name (spaces, hyphens, accents and capitals don’t matter), your domain and your domain’s name in each answer. A small model lists the other brand names in each answer, and we keep a name only if it appears in that answer word for word, so nothing is invented. Positions come from where each brand first appears.
What this is, and what it isn’t
This shows what the underlying models say when asked, through their APIs. It isn’t a capture of the live ChatGPT, Claude or Perplexity apps, which add their own search, memory and personalisation, and AI answers vary from one ask to the next. It’s a fair, repeatable reading of how the models connect your brand to your topic, which is the thing you can actually change.
Why AI names some brands and not others
A model recommends the brands it has read about most often in connection with a question: in reviews, “best of” lists, comparisons, forums and the brands’ own pages. When it searches the web, it quotes the pages that answer the question most directly right now. Both reward the same thing: clear, specific pages that answer what buyers ask, and other sites that mention you alongside your topic.
That’s the work RankSpiral does for you every day. It maps the questions your buyers search, writes the articles that answer them in your brand voice, and links your products inside them, so the answers have something of yours to draw on. See how the authority map works.
Questions people ask
Is this what ChatGPT shows my customers?
Not exactly, and we don't pretend it is. We ask the questions to the models behind the assistants, through their APIs: Claude Sonnet 5 answering from what it learned in training, and GPT-5.6 Luna with web search. That shows what the underlying models say when asked. The live ChatGPT, Claude and Perplexity apps add their own search, memory and personalisation, so what one person sees there can differ.
Why can my result change if I run it again?
AI answers aren't fixed. The same question can name a slightly different set of brands each time, the questions themselves are written fresh for every check, and web search results move. A brand that is named in most answers stays named in most answers, but a single borderline mention can come and go. Treat the score as a reading, and compare it over months rather than minutes.
How do you decide whether my brand was mentioned?
By matching text, not by asking a model. We look for your brand name as you typed it (spaces and hyphens are interchangeable, accents and capitals ignored), your domain, and your domain's name if it is distinctive. Very short names, like a two-letter brand, only count when written exactly as you typed them. A link to your site that appears only as a source citation is shown as "cited" but doesn't count as being named.
How is the position worked out?
A small model lists the brand names in each answer, and we keep a name only if it appears word for word in that answer. We then order the brands by where they first appear. Position 1 means you were the first brand the answer named; "#3 of 8" means two brands came before you, out of eight named.
We rank well on Google. Why don't the AI answers name us?
The two overlap but aren't the same. A model answering from memory names the brands it read about most often in connection with a topic: in reviews, roundups, comparisons, forums and the brand's own pages. Ranking for your own product pages doesn't always put your name next to the questions buyers ask. Pages that answer those questions directly, and mentions on other sites that cover your topic, are what close the gap.
How do I get mentioned by AI assistants?
There is no switch to flip, but the things that help are plain: publish pages that answer the exact questions your buyers ask, write fair comparison pages against the brands that are named instead of you, get included in the roundups, directories and reviews that feature those brands, and describe your brand the same way everywhere. Your report lists the steps that fit what we found for you.
What's free, and what do I get with my email?
Your first check each day is free and shows your score, two questions in full and the brand named most often. Add your email and we send you the whole report: every question and answer, every competitor named, and what to do next. Opening the report link also gives you 10 full checks a month.
Ready to rank on page one?
We will map your niche and write your first two articles, links included. Judge us on those.
3-day free trial · Two full articles on us · No credit card required