Can I check if ChatGPT knows about my website?

You can check whether it is able to: whether the crawlers are allowed, whether your content is in the HTML, whether your brand resolves as an entity. You cannot check citation share from a scan. For presence, ask a fixed set of questions on a schedule and record the rate — a single spot check tells you almost nothing, because answers vary between runs.

Why does ChatGPT cite my competitor instead of me?

Usually one of three reasons, in this order of likelihood: your page is blocked or renders client-side so it was never read; the passage that would answer the question does not answer it in its opening sentence; or the model recognises their brand as an entity and not yours. The first is binary and the cheapest to check.

Does blocking GPTBot remove me from ChatGPT?

No. GPTBot is the training crawler. Live answers are served by OAI-SearchBot and ChatGPT-User, so you can decline training use and still appear in answers — provided you have not blocked those two as well, which blanket bot rules frequently do.

Is there a paid way to appear in ChatGPT answers?

Not today. There is no advertising slot in an answer's citations, which is exactly why the structural work is worth doing now — the advantage is earned rather than bought, and it compounds while the format is still settling.

ChatGPT visibility checker

A ChatGPT visibility checker is a tool that establishes whether ChatGPT can reach, read and quote your page — not how often it actually cites you, which no scan of your HTML can determine. Any tool claiming the second from the first is inferring rather than measuring.

That distinction is the whole subject. Readiness is deterministic and under your control. Citation is an outcome that varies per model, per prompt and per week, and establishing it means querying the model repeatedly, which costs money and produces a sample rather than a fact. Conflating the two is how this category misleads people, and it is why this page exists.

Score any live URL — free, no account

What a scan can honestly tell you

Everything ChatGPT needs before it can cite you is visible in your own HTML and public records, which means it can be checked exactly and repeatedly.

The gates come first and they are binary. Does your robots.txt allow OAI-SearchBot and ChatGPT-User — the agents that fetch pages to answer live questions? Is your content in the HTML, or does it appear only after JavaScript runs, which these crawlers generally do not do? A failure at either point makes everything downstream irrelevant, and both are common and accidental.

After the gates come the gradual signals: whether each section answers its own heading directly, whether your comparisons exist as real tables, whether your JSON-LD declares who you are, and whether the brand resolves to a recognised entity. These do not guarantee citation. They are the difference between being citable and being uncitable, which is the part you can actually engineer.

Why nobody can hand you a ChatGPT visibility score

Three properties of the system make a single trustworthy number impossible, and none of them are going to change.

Retrieval is approximate, so near-ties between candidate passages resolve differently between runs. Generation is sampled rather than deterministic, so the same question asked twice can surface different sources. And the space of possible questions is unbounded — a score would have to summarise your presence across every phrasing anyone might use, which no sample can do.

This is why a score that moves when you have changed nothing is not a measurement, it is noise with a decimal point. Presence is real and worth tracking, but it has to be reported as a rate across a fixed prompt set over time, with the sample size stated.

The three-part method that actually works

Because no single number is trustworthy, instrument the three layers separately and never blend them. Each answers a different question and each has a different fix.

Readiness is the deterministic layer: scan your priority pages, record the score, track it as you make changes. Retrieval is the observable layer and almost nobody looks at it — your own server access logs show exactly which AI user agents fetched which URLs and what status they received. A run of 403s there usually means a firewall is blocking crawlers your robots.txt welcomes. Presence is the sampled layer: a fixed set of questions, run monthly, recorded as an appearance rate.

Reported together these tell a coherent story — ready, being fetched, showing up — and each one localises the problem. Reported as one composite they hide precisely the diagnostic detail you need.

Layer	How to measure it	What it tells you
Readiness	Scan the page	Whether you can be cited at all
Retrieval	Your server access logs	Whether crawlers are actually fetching you
Presence	A fixed prompt set, run monthly	Whether you are showing up, as a rate

Getting ChatGPT to reach you in the first place

The most common reason a brand is absent from ChatGPT answers is not weak content. It is a robots.txt or firewall rule nobody remembers writing.

ChatGPT reaches sites through distinct user agents that do different jobs, and they deserve separate decisions. OAI-SearchBot builds the index behind live answers. ChatGPT-User fetches a page when a user's question causes the assistant to open it. GPTBot collects training data. Blocking the third is a legitimate content position with no effect on live answers; blocking the first two removes you from those answers entirely.

Check the file, then check your logs to confirm the server agrees with it. A robots.txt that allows a crawler while a web application firewall returns 403 to the same agent is a site saying one thing and doing another, and the crawler believes the server.

Frequently asked questions

Can I check if ChatGPT knows about my website?: You can check whether it is able to: whether the crawlers are allowed, whether your content is in the HTML, whether your brand resolves as an entity. You cannot check citation share from a scan. For presence, ask a fixed set of questions on a schedule and record the rate — a single spot check tells you almost nothing, because answers vary between runs.
Why does ChatGPT cite my competitor instead of me?: Usually one of three reasons, in this order of likelihood: your page is blocked or renders client-side so it was never read; the passage that would answer the question does not answer it in its opening sentence; or the model recognises their brand as an entity and not yours. The first is binary and the cheapest to check.
Does blocking GPTBot remove me from ChatGPT?: No. GPTBot is the training crawler. Live answers are served by OAI-SearchBot and ChatGPT-User, so you can decline training use and still appear in answers — provided you have not blocked those two as well, which blanket bot rules frequently do.
Is there a paid way to appear in ChatGPT answers?: Not today. There is no advertising slot in an answer's citations, which is exactly why the structural work is worth doing now — the advantage is earned rather than bought, and it compounds while the format is still settling.

Fixing one page? Audit the whole site.

These tools work on a single page, in your browser. The full live-URL audit — SEO, GEO and entity authority — runs free with no sign-up and hands you your top fix; sign in with Google (also free) for the rest of the fixes and your saved report history.