Hey Product Hunt 👋

A single photo can end a career, win a lawsuit, or drain a bank account. The problem is that today almost no one can tell, just by looking, whether it's real.

I built Uthentic because every AI detector I tried gave me one score from one model and nothing else. So I built it differently:

→ Several independent AI engines fused after calibration, so one wrong reading never decides alone

→ A photographer's read of light, shadows, reflections and material behavior, explained in plain language, not just a percentage

→ The suspicious area circled directly on the image, with the reason for the flag

→ A real distinction between a generated image and a genuine photo that's simply been retouched, something most tools confuse

→ "Insufficient evidence" as an honest answer when the signals disagree, instead of a forced percentage

→ EXIF data, C2PA provenance and local-edit signs checked alongside the visual analysis

→ Photos AND text, each with its own engines and its own report

→ Your file is only used during processing, then deleted, never used to train models

→ Live on iPhone and Android, not just desktop browser

Every change to how the engines are combined goes through a repeatable benchmark of 10,179 tests. It's currently the most complete AI detector I know of for checking both photos and text — tools like GPTZero or ZeroGPT are solid for text alone, but stay in their own lane.

Free to try on your own photos: no credit card, no signup required for the first scan.

Would genuinely love to know what breaks it — that's the fastest way to make it better.