Answermend / acceptable use
Acceptable use
What you
can
point this at.
The four generators only transform what you type, so there’s nothing to misuse but the output. The three site checks and the AI Visibility Checker fetch a URL from our server on your behalf, and the last one also spends real money at ChatGPT and Gemini — so those three carry real limits, enforced automatically rather than reviewed by anyone. This page states them in plain words. The terms page Acceptable use states the same boundary as a contract.
Allowed
What you can do with it.
No account and no attribution required for any of it.
your own site
Check it as often as the limits below allow, and paste the output straight into it.
a client’s site
Anything you’re authorised to look at — a client’s, an employer’s, a site you’re pitching to work on.
the output
A schema block, a robots.txt rule, an llms.txt entry — it’s yours. Put it in a deliverable, charge for the work.
occasional automation
Working a check into your own process is fine, inside the limits below. It’s a loop aimed at getting past them that isn’t.
Not allowed
What gets a request refused.
Four things, and the reasoning is the same for all of them: each of these tools runs on server time, someone else’s bandwidth, or a metered call to a paid API — never on nothing.
-
Not allowed
Checking a site you have no relationship to
The three site checks and the visibility check are for a site you own, work on, or have been asked to look at — not a stranger’s business you’re curious about. A check leaves a record of our server’s IP address in their logs, and for the AI Visibility Checker it also takes their business name to ChatGPT and Gemini and keeps what came back.
-
Not allowed
Scripting or looping a check to get around a limit
Cycling addresses, automating repeats, or hammering a tool at a rate no person would use it at. The limits exist because the tools that fetch a URL are not free for us to run; getting around one doesn’t save you anything, it just spends someone else’s share of the day.
-
Not allowed
Using a generator to describe something that doesn’t exist
The generators check that what they produce is well formed. They do not check that it’s true. A schema block for a business, a product or a review that isn’t real is valid markup and a bad idea — that judgement is yours, and publishing it is misuse of the output, not of the tool.
-
Not allowed
Using any of this as one leg of an attack on another site
Working a request limit backwards, or aiming a check at a site as part of anything meant to slow it down or cost it money. That’s a misuse of this site and of the other one.
What’s actually enforced
The limits, plainly.
Read straight off the code that enforces them, not off a policy document that could drift from it. They can change without notice as the bill and the abuse require — the shape below is what’s stable, even where a number moves.
| Guard | What it limits | What you see if you hit it |
|---|---|---|
| Per-connection burst | How fast one internet connection can start a visibility check — three an hour, today. | A refusal, not a result: no score, and nothing counted as a zero. |
| Per-connection day | How many visibility checks one connection gets before the count resets — two a day, today. An office or household sharing one connection shares this allowance. | The same kind of refusal, naming when it clears. |
| The shared daily budget | The whole day’s spend on ChatGPT and Gemini calls, shared by every visitor running a check at once — not yours alone to exhaust. | Published live on the checker page itself while you use it. Once it’s gone for the day, every visitor sees the same refusal — it isn’t personal, and it resets the next day. |
| Repeat checks | Whether the domain you asked about was checked recently by anyone. | The existing result, handed back immediately and not charged against your allowance — the fastest and cheapest thing this system can do for you. |
| The three site checks | Crawler Checker, Readiness Score and Render Check sit behind a lighter, general request limit, because each one only fetches your page once. | A refusal if they’re scripted or run in a loop. |
Every refusal on this page follows the same rule the measurement itself does: an unrun check is not a zero. A limit hit mid-request means nothing was asked of any engine on your behalf that time — there is no score, no partial result, and nothing here is dressed up as one. What “not measured” means for a completed check, rather than a refused one, is on the docs page built for exactly that question.
This is the plain-language version.
The terms page states the same boundary as the binding contract. If a refusal caught you and you think it shouldn’t have, write in — a person reads it.