AI Watermark Checker

What this tool actually checks

It reads the structured records embedded in the file: C2PA manifests, XMP packets, EXIF blocks, IPTC records, and the text chunks that image generators write into PNGs. These are the places where an AI tool leaves its name, and where a camera leaves your location.

It reports what it finds and, just as importantly, what it cannot determine. A signal that no browser-side tool can measure is labelled "Unable to verify" rather than "Not detected", because those two statements mean very different things.

Two different kinds of watermark

Most things people call an "AI watermark" are metadata: a name in a field, a manifest attached to the file. Metadata is separable from the picture, so it can be read and removed cleanly.

A smaller category is embedded in the pixels themselves. Google's SynthID works this way. It survives metadata stripping, it survives re-saving, and detecting it requires the vendor's own detector. No tool that runs in your browser can tell you whether it is there.

A checker that blurs this distinction is not being helpful. This one keeps them separate.

Privacy

The scan runs entirely inside your browser. Your file is read from disk into memory, inspected there, and never uploaded — you can disconnect from the internet after this page loads and everything still works.