AI Visibility Checker
that ends in a fix.
Type your domain. Penwell asks ChatGPT, Claude, Gemini and Perplexity whether they name you, checks whether their crawlers can read your site at all, and writes the file that fixes it. Free, no signup.
Works on any site that answers, including one that isn't yours.
What the check reads
- Whether the crawlers get in
- GPTBot, ClaudeBot, PerplexityBot, Google-Extended and CCBot, checked against your real robots.txt. A crawler turned away at the door never sees anything else you did, so this is the half that gets a score.
- Whether the models name you
- One buyer-intent question, put to ChatGPT, Claude, Gemini and Perplexity with search switched on, and the answer quoted back the way it came.
- The pages they read to answer
- The urls the models actually opened. It is the practical version of the finding — if your name is not on those pages, that is where it has to get to.
- An llms.txt built from your own site
- Not a template with your domain dropped in. It is generated from your page title, your meta description and the urls in your sitemap, and you can paste it as printed.
What it refuses to tell you
Most AI visibility tools open with a score out of 100 and a chart of it moving. Almost none of them can tell you how many times they asked, and the honest answer is usually once — which makes the number a measurement of the sampling rather than of you. A product named this week and missing next week has very likely not lost anything at all.
So there is no score here for whether the models name you, and there never will be. You get the count, the answer quoted the way it came back, the other products it named instead, and the pages it read to decide. When an engine errors or times out, that shows as a missed reading rather than as a no, because a gap is not an absence.
The crawlability half is the opposite: robots.txt either allows the AI agents or it does not, llms.txt either exists or it does not, and the same site checked twice gives the same answer. That half is scored, it costs us nothing to run, and it is the one you can fix this afternoon.
Questions
- Do I have to sign up?
- No. There is no account, no card and no permission screen. Everything read here is already public — the same files and pages an AI crawler fetches. We cache the report for a day, so re-opening it is free.
- Which models do you ask?
- ChatGPT, Claude, Google Gemini and Perplexity — 4, each asked once, each with web search on. Copilot and Google's AI Overviews are not included: neither has an API, and covering them would mean scraping a search page and calling the result a measurement.
- Why is there no AI visibility score?
- Because it would be a random number with a chart on it. Ask a model the same question twice and it does not answer the same way, so a percentage off a single sample measures the sampling. We show the count, quote the answer and say outright that tomorrow may differ. The crawl half is deterministic, so that one does get a score.
- What is llms.txt, and do I need one?
- It is a plain text file at the root of your site that tells a model what the site is and which pages matter — the same idea as robots.txt, aimed at readers rather than crawlers. Nothing breaks without one; it is the cheapest way to control how you get summarised. This page generates yours from what it just read.
- Can I check a competitor's site?
- Yes. Any site that answers a plain request works, and the question we put to the models is one anybody could type into a chat box.
Looking for something else? All the free tools →
One check is today. Penwell is the one that keeps asking.
Continue with X