Free tool

Can ChatGPT even read your website?

Nine sites out of ten block at least one AI agent without knowing it, or publish content those agents never see. This analysis tells you in thirty seconds, point by point, and you can verify every finding yourself.

GPTBot OAI-SearchBot PerplexityBot ClaudeBot Google-Extended CCBot

Immediate result, no sign-up.

Reading robots.txt
Analysing the content
Looking up the entity
Structure and freshness

What this scanner measures, and what it does not

It measures what decides whether an answer engine can read and cite you: its agents access to your site, whether your content really sits in the HTML, whether your brand exists as a named entity, how your pages are structured and how fresh they are. Every point is observed, never estimated.

It does not claim to tell you whether ChatGPT cites you today on a given question. That is measured over time, question by question, and that is what Pulsar Ranking does continuously. This scanner answers the question that comes first: is your site even fit to be cited.

How to read your score

The five families do not carry equal weight. Agent access accounts for more than a third, because a blocked agent reads nothing and no other optimisation makes up for it. Readability comes next: content rendered by JavaScript does not exist for an agent reading raw HTML.

A score above 80 means nothing stands in the way of being cited. Between 55 and 80, points are being lost without anything being blocked. Below that, a structural obstacle keeps your brand from appearing, whatever editorial work you put in elsewhere.

Frequently asked questions

Does blocking AI agents protect my content?
Mostly it stops you being cited. Blocking GPTBot does not stop your content circulating; it stops ChatGPT naming you as the source when it does. The decision is commercial, not technical: would you rather be cited, or invisible.
My site runs on React or Vue, is that a problem?
It depends on the rendering. If pages are server-rendered, the content sits in the HTML and all is well. If they are rendered in the browser, agents receive an empty page. The line about text in the HTML source tells you precisely.
How often should I run this again?
After every site change, and once a quarter otherwise. Agent lists move fast: an agent that did not exist last year may now be caught by a generic rule in your robots.txt.
Do you keep the sites analysed?
The analysed address is kept to measure use of the tool. Your e-mail is only stored if you ask for the report, and is used for nothing else.