Check yours
Is it a scam? Check a message, call, advert or website.
Paste a text, email or link, describe a phone call, or upload a screenshot. We look for the warning signs scammers use, including AI tricks such as cloned voices and fake celebrity videos, check when any website in it was set up, and tell you in plain words what to do next. It can never promise that something is safe, and it says plainly what it could not confirm.
Free, no account needed. Automated results are a starting point, not a published editorial verdict.
Check something β free evidence checker
Choose what you received. It tells the check which scam tricks to look for. It never decides the answer.
Never paste passwords, one-time codes or card numbers. Remove your own account details first.
Video is securely processed using specialist third-party infrastructure. Selected transcript and representative-frame evidence may be analysed by AI services to produce the Full Evidence Report. Visual analysis is sampled, not frame-by-frame forensic examination.
Free checks examine browser-readable evidence first. Document contents are analysed in the Full Evidence Report (coming soon), not inferred from a filename.
No account required.
Results guide further checking; they do not replace reading the cited evidence in context.
Before you check
Good to know.
Can it tell me that something is safe?
No. It looks for warning signs and checks what public sources say, then tells you plainly what it found and what it could not confirm. “We did not spot the usual scam signs” never means something is safe. If money or bank details are involved, contact the organisation using details you already trust, not the ones in the message.
Can it check a phone call or a video?
It cannot listen to audio or watch video. Describe what was said, or paste the advert’s words or link, and it checks those against the patterns used in scams. Cloned voices and fake celebrity videos are now common: a familiar voice or face is not proof. Hang up and call back on a number you already trust, and agree a family code word for emergencies.
What happens to what I send?
Your text, link or screenshot is sent to our AI provider, OpenAI, so the words can be read and checked against public sources. We ask OpenAI not to store it and we do not keep it after the check. Please remove passwords, one-time codes, card numbers and your own account details first. Our privacy policy has the detail.
Is there a limit?
Each person can run up to ten checks every ten minutes, and there is a daily limit across the whole site. If it has been reached, the checker opens again the next day.
Can an AI detector prove something was made by AI?
No. Detector scores are guesses. We never rely on them for an answer; see how we check.
I have already paid or shared my details. What now?
Contact your bank straight away using the number on your card or its official website, then report it to the police. Do not wait for, or rely on, a result from this checker. Our scams page lists where to report it.
Today’s Real or AI? round: five pictures, about a minute.
Scams going round, once a week
One short email on Fridays: the new scams we have checked and what to do about them. Unsubscribe any time.