Before a secondhand deal (Depop, Vinted, Carousell…), check the listing and seller. You can also test whether listing photos are AI-made or stolen.
Free with no sign-up, and the analysis runs inside your browser — files never leave your device. On our public gate (30 real + 83 AI images) it measured 94.7% at original quality and 87.6% on recompressed copies, with zero real photos confidently mislabelled as AI. Treat it as an indicator, not proof: the newest commercial generators evade every detector.
Secondhand-deal scams are decided by the payment path, not the item. Inside platform-protected payment you are mostly safe; the moment you leave it, you are not. You can also test whether listing photos are AI-generated or stolen.
A scam listing has no item, so its photos come from somewhere else — store product shots, other listings, or increasingly AI generation. Ask the seller for a photo with a note showing today's date. With a real item that is a ten-second favor; without one, they dodge. Run received photos through the home checker for AI signals and reverse image search for theft.
A price clearly below market usually means stolen goods, defects or a scam. The more elaborate the story about why it is cheap, the more careful you should be — stories cannot be verified, but the payment path can.
Marketplace fraud usually starts with bait: clearly below market price, with a reason that comes from a fixed set — moving house in a hurry, an unwanted gift, leaving the country. None of those is inherently a problem, but combined with the price they are reason to check.
Photos are the stronger clue. Manufacturer images or pictures lifted from other listings are common. Run the listing photo through a reverse image search to see whether it appeared elsewhere first, and ask for a photo of the item beside a note with today's date. A genuine seller sends one within minutes.
Sellers using AI-generated images have started to appear too. If the product shots look unusually clean or the background is strange, run them through the image check on this site. A generated-image result is not proof of fraud in itself, only grounds to insist on real photographs.
However convincing the photos and price, one question usually settles the matter. If escrow is refused in favour of a direct transfer, you can stop there. The seller's reasoning about fees is always the same, and that fee is precisely the cost of avoiding this situation.
Check the account holder too. If the name on the bank account differs from the name the seller gave, you may be dealing with a rented account or a third party in the middle. And running the account number and phone through a fraud-report lookup before paying takes seconds.
Meeting in person is not a cure-all either. It is good for inspecting the item but carries its own risk with large cash amounts. Meet in daylight somewhere busy, and with electronics power the device up on the spot to check it is not locked or still registered to someone else.
The moment just before a deal closes is the most dangerous. You already want the item, and that is exactly when the seller applies time pressure. So make the final check a fixed sequence you run mechanically.
Look up the account number and phone for fraud reports. Ask for the photo with today's date beside the item — evasion or irritation at this request is itself the answer. And start escrow inside the platform app rather than through any link you were sent.
Keep the conversation. What a report needs is the full exchange and the transfer record, so leave it in place for a few days even after a good deal. A suggestion to move to an outside messenger is often an attempt to remove exactly that record.
The listing images looked natural, right down to wear marks, so they were clearly photographs of a real item. A reverse image search returned nothing. People do take photos from someone else's listing, so a genuine photo does not mean the seller has the item. Asking for a shot of it beside a note with today's date resolves this within minutes — and evasion or anger at the request is itself the answer.
Asked to use the platform’s protected payment, the seller complained about the fee and pushed for a bank transfer — and the listing photos turned up in another listing via reverse image search. Payment route and photo provenance fire together. Marketplace fraud is almost always decided by whether the money left the platform.
Asked for a new photo of the item next to a note with today’s date, the seller sent one within minutes, with lighting and background continuous with the original listing. This request is the cheapest possible filter for stolen photos. Only someone holding the item can pass it, and photo thieves go quiet at exactly this point.
No results feels like a clean bill of health. But photos outside search indexes — private accounts, recent shots, other platforms — never appear, and AI-generated images have no original to find. "No results" is missing information, not innocence. Pair it with an AI-probability check on the photo and a request for fresh pictures of the item.
Losing the full price to save a small fee is the basic shape of the scam. Treat the fee as insurance. A genuine seller has no reason to refuse protected payment.
Good for checking the item, but large cash deals carry their own risk. Meet in daylight at a public spot, and test high-value electronics on the spot before paying.
Yes. Sending a link to a page imitating an escrow service is a known method. Run any link you receive through the link check, and above all start the payment flow inside the platform app yourself. The moment you leave for an external link, it is not escrow.