0145 %
Access
Your robots.txt, your WAF or a forgotten noindex may be shutting the door. If they can't get in, nothing else matters.
- robots.txt per agent
- Headers and noindex
- Cloudflare and anti-bot
Technical check · Free · Up to 12 pages
Can AI get into your site and read it?
We crawl several pages of your site and analyse them exactly the way ChatGPT, Perplexity, Claude and Google's AI Overviews do: without running JavaScript. We tell you what they see, what they miss and on which pages.
Check how visible your site is to ChatGPT, Perplexity and Google AI Overviews — free and with no sign-up.
- No sign-up
- No email needed to see the result
- Result in seconds
We check access for 19 crawlers
Agents checked: GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-User, Claude-SearchBot, PerplexityBot, Perplexity-User, Googlebot, Google-Extended, Bingbot, Applebot, Applebot-Extended, DuckAssistBot, MistralAI-User, CCBot, meta-externalagent, Amazonbot, Bytespider.
- 1.284
- Sites analysed
- 29
- Technical checks
- 19
- AI crawlers
How an assistant reaches you
First the index, then a live read. We check both.
Assistants answer first from the snippets in the search index, and only then fetch your page live: fast and without JavaScript. A site can pass the first step and fail the second without anyone noticing.
Structure and Access feed the index.
- Page title
- Meta description
- Indexability
- Sitemap
- Canonical
Access and Readability decide the live read.
- robots.txt per bot
- WAF / 403
- Content without JavaScript
- H1
- TTFB
- HTML size
Three pillars
Get in, read, understand. In that order.
The weights aren't equal on purpose. A flawless site that lets nobody in is still invisible.
- Access
- 45 %
- Readability
- 35 %
- Structure
- 20 %
0235 %
Readability
No AI crawler that matters runs JavaScript. We check what's left of your page once it's switched off.
- Content in the HTML
- Empty shells
- noscript fallback
0320 %
Structure
What lets a machine work out what your site is about, and which entity is behind it, without guessing.
- JSON-LD and its validity
- Canonical and sitemap
- Titles and language
What we don't do
We're not going to tell you we can predict whether ChatGPT will cite you.
Nobody can, and anyone promising it is selling smoke. The available data suggests that even the most strongly correlated external signal explains only a small fraction of why a model recommends a brand. What is measurable, stable and actionable is the prerequisite: that they can get in and read you.
How we calculate the scoreThe blind spot
Your site, seen by an AI crawler
Plenty of modern sites build their content in the browser. A person sees it. The crawler gets an empty shell.
<!doctype html>
<html lang="es">
<head>
<title>Inicio</title>
</head>
<body>
<div id="__next"></div>
<script src="/main.js"></script>
</body>
</html>Zero words of content. Nothing to cite.
AI SEO
AI SEO and ChatGPT SEO: what it is and where to start
You keep hearing «AI SEO» or «SEO for ChatGPT», and the two don't always mean the same thing. Here's the plain version.
«AI SEO» can mean two different things
- 01
Using AI tools to do traditional SEO better
A language model that researches keywords or drafts copy, for instance. That's not what this page is about.
- 02
Optimising your site so that AI systems themselves can find you, read you and cite you
This is what we do here, and what we mean by «SEO for ChatGPT» in the rest of the page.
Where do you start?
The same place as any serious SEO project: making sure there are no technical barriers before you worry about anything else. AI SEO works the same way, with one important difference — the barriers aren't the same ones.
A site that ranks well on Google can be completely closed to ChatGPT without anyone noticing, because AI crawlers follow different access rules from Google's.
Our free check does exactly that: it tells you, in seconds, whether your site has any of those barriers.
Glossary
The terms you're going to run into
Before we go on, let's sort out the vocabulary. You don't need to be technical to follow it: these are simple ideas with names that sound harder than they are.
What is GEO?
GEO (Generative Engine Optimization) is the equivalent of SEO, but aimed at being found and cited by AI assistants —ChatGPT, Perplexity, Gemini, Claude— instead of Google, or as well as it. Classic SEO works to place you in a list of results; GEO, to let a model get in, read you and cite you inside an answer. The groundwork is similar; the access requirements are not.
What is an AI crawler?
An automated program that AI companies use to visit pages and read what is on them, much as Googlebot does for Google but with a different user-agent: GPTBot (OpenAI), ClaudeBot (Anthropic) or PerplexityBot (Perplexity). If your site will not let it in, that assistant can never read you or cite you. Blocking them is a legitimate choice; doing it by accident is not.
What is robots.txt?
A text file living at the root of your site (yoursite.com/robots.txt) telling each crawler where it may go. It is the first door: if an AI bot is blocked here, it never gets to see anything else, however well optimised the rest may be. One bad line shuts all of them out, and it is the first thing we check. Most sites that fail here never knew the file was there.
What is llms.txt?
A file proposed recently to tell models which content on your site matters most, as a sort of guided summary. The idea is interesting, but let us be honest: there is no public evidence today that the main crawlers take it into account. Having one does no harm; treating it as a miracle cure does. We report whether you have one, and leave it at that.
What is structured data?
A block of code, invisible to a human visitor, that states explicitly what each thing on your page is: «this is an article», «this is a frequently asked question», «this is a company called X». That way neither Google nor the models have to infer it from the visible text. It is the difference between declaring and hoping.
Why does it matter if my site uses a lot of JavaScript?
Most AI crawlers do not run JavaScript: they read the HTML as it arrives, without opening it in a browser. If the main content is built with JavaScript —common in React, Vue or Angular— what they see may be an almost empty page, even though a person sees it perfectly. It is the most common problem and the least visible.
What is indexability?
A page's ability to be stored and shown by a search engine or an AI assistant. A page can be reachable and still not indexable: one noindex tag added by accident, or a misconfigured redirect that stops it being stored. Reachable and indexable are not the same thing, and the difference is easy to miss.
Guides
How to fix the most common problems
Letting AI in through your robots.txt
- Open yoursite.com/robots.txt in your browser. If it doesn't exist, your server probably allows everything by default — but it's worth creating one explicitly.
- Look for «User-agent:» lines followed by «Disallow: /». If you see that next to GPTBot, ClaudeBot, PerplexityBot or Google-Extended, those bots are shut out.
- Change «Disallow: /» to «Allow: /» on those user-agents' lines, or delete the specific block if you don't need to restrict anything.
- Save the file at the root of the domain and check again.
Working out whether your content leans too hard on JavaScript
- Right-click your page and choose «View page source», not «Inspect»: the latter is already rendered by the browser.
- If you see little more than an empty container and hardly any real text, your content depends on JavaScript to exist at all.
- An AI crawler that doesn't run JavaScript sees exactly that: an almost empty page.
- The fix is server-side rendering (SSR) or static generation (SSG), so that the HTML you send already carries the content.
Adding basic structured data (JSON-LD)
- Work out what kind of page it is: an article? a frequently asked questions page? your company profile? Each type has its own schema.
- Generate the matching block of code. There are free online generators that don't require you to write any code.
- Paste the block inside the page's <head>, wrapped in a <script type="application/ld+json"> tag.
- Check that the result is valid JSON: a single syntax error invalidates the whole block.
Checking whether your anti-bot protection blocks AI by mistake
- Check whether you have Cloudflare, Sucuri or another WAF in front of your site.
- Review your «Bot Fight Mode» settings or your anti-bot rules: badly tuned, they block legitimate AI crawlers too.
- Add explicit exceptions for the AI user-agents you want to let through.
- Analyse your site again: if the challenge is still there, we detect it and tell you.
Frequently asked questions
Frequently asked questions
How do I know whether ChatGPT can read my site?
The quickest way is to look at your robots.txt and see whether any AI user-agent (GPTBot, ClaudeBot, PerplexityBot, Google-Extended) is blocked, then check whether your main content depends on JavaScript. Our tool runs both checks at once, in seconds and with no sign-up.
Why doesn't ChatGPT recommend my company when I rank well on Google?
Ranking well on Google is no guarantee that an AI model can read you: they are different systems, with different access rules. It is common for a site that ranks well in traditional search to have technical blocks it knows nothing about, blocks that shut AI crawlers out completely.
Is there any way to guarantee that ChatGPT cites me?
No, and anyone who promises it isn't being honest. The public data available shows that even the external signal most strongly correlated with being recommended by a model explains only a small fraction of the outcome: most of the decision happens inside the model, where it can be neither observed nor predicted from outside. What can be measured and fixed is the prerequisite: that the crawler can get in and read your content.
How is this check different from a normal SEO audit?
A traditional SEO audit is about rankings: keywords, links, content. This check is only about an earlier and more basic problem: whether AI systems can technically reach your site and read what's on it. It doesn't replace SEO; it's an extra requirement that many traditional audits never test.
Does the free check have any hidden costs?
No. The quick analysis needs no sign-up, no email and no card. If you want a fuller report with detailed, prioritised recommendations, we offer an extended version — but the initial check is free, with no strings.
How often should I repeat this analysis?
After any meaningful change to your site — a redesign, a change of CMS, a new security setup — and, at the very least, every few months. Crawler access rules and your own infrastructure settings can change without you noticing.
Does it work for any kind of site?
Yes. The technical checks on access, readability and structure apply to any site, whether it's an online shop, a blog or a services company. What changes is how much each point weighs, depending on the kind of content you have.
What matters more: showing up on Google or being readable by AI?
They aren't mutually exclusive, and they are increasingly the same thing. More and more searches start or end in an AI assistant rather than in a traditional list of results. Ignoring AI accessibility is, more and more, leaving a door shut without knowing it.
Check it on your own site
Instant result, up to 12 pages. No sign-up and no email to hand over.