Hi everyone. I built PEA in Japan after seeing how often fabricated screenshots and edited photos turn up in everyday disputes: chat logs in a breakup, a "damaged item" photo in a marketplace refund, a dented car after a parking-lot bump. What it does: you upload a photo or screenshot, and PEA runs eight independent analyses and gives a grade from 1 to 10 with the reasons listed, confirmed facts on one side and concerns on the other. If there isn't enough to go on (a still from a video, a photo of a screen), it says "Not assessed" rather than guessing. What it doesn't do: it doesn't give a percentage, it doesn't promise a court will accept anything, and it can't prove who took a photo or when. It's a first check and a way to preserve the file: a certificate fixes the SHA-256 hash, signs it with Ed25519 and adds a third-party RFC 3161 timestamp, and anyone can verify it by number on a public page. Privacy: staff never see your images unless you ask for a human review and agree to share. The free check works in English and needs no account (pea.luminasparkdesign.com/en/check). Account and certificate screens are still in Japanese; browser translation works well on them, and a full English version is in early access. I'd love to hear what you'd need from it.

Forum Threads

p/pea-evidence-authenticator

•

3d ago

Would you rather get one AI-probability number, or the reasons?

We launch PEA on Oct 17. Before that, a question for anyone who has used AI image detectors.

Most tools give one number ("87% AI"). PEA lists what it confirmed and what raised concern instead, and refuses to grade images it can't judge, like a still from a video or a photo of a screen.

Which would you trust more, and what would you actually check with a tool like this?

View all

PEA – Evidence Authenticator

Check if a photo or screenshot is AI-made, with reasons

Check if a photo or screenshot is AI-made, with reasons

Forum Threads

Would you rather get one AI-probability number, or the reasons?

Forum Threads

Would you rather get one AI-probability number, or the reasons?