Your client's site will break silently.
When ChatGPT, Claude or Perplexity can no longer read a website, nothing goes red. No error, no warning, no drop in Google rankings. It just quietly stops being recommended. We watch for it and tell you the same day.
Anyone can tell you what your robots.txt says this morning. Only something that has been watching can tell you GPTBot used to come every day and has not been back in two weeks.
A real check against 19 AI crawlers and opt-out rules. We will show you a sample of what we find, in about 5 seconds. No signup needed to look.
Your result is private. You get a link you can send to a client if you want to; the address on its own will not open it.
Not many visitors. Extremely good ones.
We will be straight with you, because the honest version is more convincing than the sales pitch.
Google still sends 87.6%. Anyone telling you AI has replaced search is selling something.
They arrive already convinced. One B2B study: 0.5% of visits drove 12.1% of signups.
Claude referrals grew 320% in a year. Gemini 231%. This is the direction, not the destination.
So the real argument is the imbalance
Being readable by AI costs you nothing. Being invisible costs you the entire channel — and you will not find out for months, because nothing reports it. That is a bad trade to lose by accident.
Figures from independent 2026 traffic studies. We would rather show you the small number than pretend it is a big one.
Nobody breaks it on purpose
Every one of these happens without anyone deciding it should. That is exactly why nobody catches it.
A security plugin updates itself
WordPress security plugins ship new "block bad bots" defaults all the time. AI crawlers get swept up with the scrapers, and nothing tells you.
Your host adds a firewall rule
Hosting companies block crawler traffic to cut bandwidth costs. Nobody emails the site owner about it.
Someone edits robots.txt
One line copied from a template in 2023 can still be quietly turning away every AI crawler today.
The site gets rebuilt
A modern JavaScript rebuild looks beautiful to people and arrives as an empty shell to crawlers that do not run JavaScript.
Set it up once, then forget it exists
Until the day it emails you — which is the only day it matters.
Add a domain
Prove you control it with a DNS record, a file, or a meta tag. Nothing to install on the site itself.
We check every six hours (daily on free)
All 19 AI crawlers and opt-out rules, plus robots.txt, JavaScript rendering, llms.txt and your page basics — and we keep every result, which is what lets us tell you what changed rather than only what is true today.
You hear only about changes
Silence when things are fine. An email the moment something breaks — or the moment a crawler that used to visit stops — with the exact line to change.
Checked every 6 hours on a paid plan, once a day on free. You only hear from us when something changes.
A user-agent is a claim. We deal in proof.
Your logs say GPTBot. That proves somebody sent a request with GPTBot written on it — and anyone can. If you see a thousand of them, you do not know how many were really OpenAI. Maybe all. Maybe none. Every AI traffic number you have rests on that difference.
So we rank our evidence instead of flattening it into a yes or no. When a bot-verifying CDN sits in front of a site and we honestly cannot tell, we say "we cannot confirm this from outside" rather than guess. Every other checker calls that a block and sends you chasing a problem that may not exist.
A real crawler reached your site and its address matched the range its operator publishes. Proof, not inference.
robots.txt permits it and our live request got through. Both agree.
Proven by robots.txt itself, or a server that refused the crawler by name.
A bot-verifying CDN challenged us. Here is how to settle it in 30 seconds.
And the one thing no checker can ever tell you
GPTBot fetched this site every day for a month, then stopped fourteen days ago. You cannot arrive at that by testing a site today, however good the test. It only exists if something was already watching and kept the record. A crawler that had a habit and broke it is usually a block nobody knew they had — and it is the single most useful sentence this product produces.
A monitor that cries wolf gets switched off. We would rather be useful for years.
19 crawlers and opt-out rules, watched continuously
When a new AI crawler appears, we add it for everyone. You do not have to keep up with any of this.
Find out where your client sites stand
Free for your first site, forever. No card. If nothing is wrong, we will tell you that too and leave you alone.
Set up in 2 minutes · Cancel any time · We only store your email and the addresses you ask us to watch