Fake Payment Screenshot Detector

The scam, and why it works

It nearly always runs the same way. A buyer, tenant or client says the money is sent, and produces a screenshot to prove it. Bank app, PayPal, Venmo, Wise, Revolut, a mobile banking SMS. The amount is right, your name is on it, the timestamp is minutes ago, and the balance looks plausible.

The money never arrives. By the time you notice, the goods are gone or the deposit is spent.

The screenshot works because it is the one piece of evidence people accept without question. Nobody expects a bank transfer confirmation to be fake, and the fake does not have to be good. It has to survive ten seconds on a phone screen while you are distracted and want the deal to be real.

What editing leaves behind

Editing a number in a screenshot is a compression event, and compression events are visible.

A screenshot is one flat JPEG or PNG with a single uniform compression history. When someone edits an amount and re-saves, the edited block gets compressed twice while the rest of the image is compressed once. Error Level Analysis makes that difference visible as a brighter patch in an otherwise even field. It is the clearest signal there is on this kind of image, because the background is so uniform that any inconsistency stands out.

Text is the other tell. Reproducing a bank app's font, weight, spacing and antialiasing well enough to survive pixel-level inspection is much harder than it looks. Edited digits are commonly a fraction off in kerning, sit a pixel high, or carry slightly different edge softness from the characters beside them.

Then there is metadata. A genuine screenshot from a phone carries a coherent set of properties: device, dimensions matching a real screen, a creation time. Open it in an editor and save, and that record is rewritten or stripped. Absence proves nothing on its own, since messaging apps strip metadata too, but it is one more signal in the stack.

Checks worth doing yourself, first

Before running anything, look at the screenshot properly. Some fakes fail on their own.

Does the status bar make sense? Battery, signal and clock should be consistent with a real device, and the clock should be plausible against the transaction time. Is the transaction time itself possible - a bank transfer confirmed at 3am on a Sunday is worth a question.

Do the numbers add up? If a balance is shown, does it reconcile with the amount sent? Scammers frequently edit the transfer amount and forget the running balance beneath it.

Is the app's own layout right? Fonts, spacing and iconography are the hardest thing to fake convincingly, and people who spend all day in their own banking app often notice something is off before they can explain what.

And the one that settles it regardless of any of the above: check your own account. A screenshot is a picture of a claim, not a payment.

What a scan can and cannot tell you

FauxLens tells you whether the image was edited. It cannot tell you whether the payment happened.

A screenshot can be entirely unedited and still worthless - taken from a real transfer that was later reversed, from a different recipient, from a test transaction, or simply of a payment that was cancelled before it settled. Equally, a screenshot can be cropped or resized innocently and still show compression irregularities.

So read the result as one input. A clean scan plus money in your account is confirmation. A flagged scan is a reason to stop and verify directly with your bank rather than to accuse anyone. The evidence chain shows which signal fired and why, so you can judge for yourself rather than trusting a verdict.