GEO checker: can AI read your website?

We run 26 technical checks in a few seconds: whether ChatGPT, Claude, Perplexity and Google can get into your site, whether they understand who you are, and what they need to cite you.

Free, no sign-up. We only read public pages, like a search engine would.

What the checker reviews

Five areas and 26 checks. They're the technical requirements for an AI assistant to find your site, read it and trust it.

Crawler access

Whether ChatGPT, Claude, Perplexity and Google can get in.

AI assistants that answer live
These are the crawlers ChatGPT, Claude and Perplexity use to search and read pages while they answer. Block them and they can't cite or link you.
Google and Bing
Google's AI Overviews come from Google's index, and Copilot and part of ChatGPT search rely on Bing's. Without these two there's no GEO.
Firewall and CDN
Many CDNs, Cloudflare included, block AI crawlers by default even when robots.txt allows them. From the outside everything looks fine, but the AI gets an error.
Permission to index and quote
noindex, nosnippet or max-snippet:0 tell search engines not to show the page or not to quote its text. Google applies them to its AI Overviews too.
Training crawlers
They decide whether your content ends up in the data models are trained on. Blocking them is legitimate, but then the model will know less about your brand when it answers without searching.
robots.txt
It's the first file every crawler reads. If it fails with a server error, Google and others stop crawling the whole site as a precaution.

Readable content

Whether the text is in the HTML and makes sense without JavaScript.

Text in the HTML
Most AI crawlers don't run JavaScript: they read the HTML as the server sends it. If the text is painted later in the browser, to them the page is empty.
Page title
It's the first thing search engines and assistants read to know what the page is about, and often the link text when they cite you.
Meta description
It doesn't rank on its own, but assistants use it as the page summary when deciding whether the page is worth reading.
Main heading (H1)
The H1 tells any reader, human or machine, what the page is about. With none, or several, the topic gets blurry.
Section structure
AIs split pages into sections to keep the passage that answers the question. Without subheadings the text is one block that's hard to quote.
Answers to real questions
People ask AI in natural language. Pages with sections that answer specific questions match what the assistant is looking for to cite.
Lists and tables
Steps, comparisons and data in lists or tables are the passages an AI extracts and reproduces most easily.
Declared language
The lang attribute tells crawlers what language the page is in, and helps you show up for people asking in that language.

Structured data

Whether your site states clearly who you are and what you offer.

Structured data (JSON-LD)
Schema.org in JSON-LD is the standard way to tell a machine, unambiguously, what each thing on your site is: your company, your services, your FAQs.
Your company profile
An Organization block with name, site, logo and profiles helps AI understand your brand is a specific entity and not confuse it with another of the same name.
Linked profiles (sameAs)
Assistants cross-check what your site says against other sources. Linking your official profiles (LinkedIn, Google Business, Wikipedia, industry directories) ties those mentions to the same company.
Page content markup
Beyond who you are, it helps to mark what each page contains: a service, an article, a product, FAQs or the breadcrumb trail.
Share preview (Open Graph)
og: tags control how your page looks when someone shares it or an assistant links it, and several tools read them as a summary.

Trust signals

What makes an AI treat you as a reliable source.

HTTPS
Search engines and assistants distrust sites without a secure connection, and some won't link them.
Canonical URL
The canonical tag says which version of the page is the right one. If it's missing or points elsewhere, signals split across duplicates or the AI cites the wrong address.
Who's behind it
AI and Google give more weight to sources that make clear who's behind them: a company or team page, contact details, an address.
Last updated date
Given two similar sources, assistants prefer the more recent one, especially for prices, regulations or comparisons. Without a date they can't tell.

Files and performance

Sitemap, llms.txt and how fast your server responds.

XML sitemap
The sitemap is the list of pages you want crawled. It helps search engines and AI crawlers find everything, including new pages.
llms.txt
A recent proposal: a Markdown summary of your site written for language models. No major search engine has confirmed using it for ranking, but it's cheap and some AI tools already read it.
Response time
AI crawlers work with short timeouts. If your server is slow they may give up before reading the page.

What GEO is, and what this tool doesn't measure

GEO (Generative Engine Optimization) is the work of getting your company into the answers of ChatGPT, Claude, Perplexity, Gemini or Google's AI Overviews. It doesn't replace SEO: those answers come from pages search engines have already crawled and indexed.

This checker looks at the technical base: whether crawlers can get in, whether the content is in the HTML, whether your site states who you are with structured data, and whether it shows trust signals. That's the part that depends only on you and is quick to fix.

What it doesn't measure: whether you're actually cited today, your domain's authority, or what other sites say about you. A technically perfect site can still go unmentioned if its content isn't useful or nobody links to it. That takes content and reputation work.

How we do SEO and GEO

Frequently asked questions

Yes. Check as many sites as you like without signing up. The full report is on the page; we only ask for your email if you want it sent to you.

SEO works to get your site into Google and Bing results. GEO works to get AI assistants to cite or recommend you when they answer. They share a technical base (crawling, indexing, structured data), but GEO puts more weight on content being easy to extract and on your brand being a clear entity.

It depends on what you sell. There are two kinds: search crawlers (OAI-SearchBot, Claude-SearchBot, PerplexityBot) that read your site to cite it in answers, and training crawlers (GPTBot, ClaudeBot, Google-Extended) that use it to train models. If you want to be cited, don't block the search ones. You can block training crawlers if your content is your product.

It's a recent proposal and no major search engine has confirmed using it for ranking. Some AI tools do read it. It's cheap to add, so we count it as an improvement, not a failure.

Because the block can live in your firewall or CDN, not in robots.txt. Cloudflare, for example, blocks AI crawlers by default on many sites. That's why we request your page as GPTBot, ClaudeBot and PerplexityBot and compare the response with a browser's.

No. The score measures whether AI can read and understand your site. Being cited also depends on your content answering what people ask and on other sources talking about you. Nobody can guarantee a mention.

We read the site once, like a search engine would, and keep the result for a few minutes so we don't repeat the check. We only use your email to send the report and, if you're interested, talk about it. Details are in the privacy policy.