Phones and cameras embed EXIF metadata in photos, including precise GPS coordinates, timestamps, and device details. Sharing an unprocessed photo can reveal where it was taken, down to a few meters. This tool shows everything embedded in the file, then removes EXIF/XMP/IPTC segments and C2PA manifests from JPEG files and metadata chunks from PNG files without recompressing the image, and re-scans the cleaned file to prove nothing remains.
Note: photos exported by messaging and social apps are often stripped already, so a clean scan of a shared photo is normal. Test with an original straight from your camera roll.
Images increasingly carry a record of how they were made. C2PA Content Credentials are a signed manifest, stored in a JPEG APP11/JUMBF segment or in a PNG caBX chunk, listing the tool that produced the file, the edits applied to it, and the IPTC digital source type, which is where a generative model shows up as trainedAlgorithmicMedia. The panel reads that manifest and reports what it claims, including the names on the signing certificates, but it does not check the signature: use the Content Credentials verifier for that.
The same claims often appear in plain metadata as well, through the IPTC digital source type, a generator name in Software or CreatorTool, or the prompt and sampler settings that PNG generators write into text chunks. Those are ordinary fields and disappear with the rest of the metadata when you download the cleaned image.
Invisible watermarks are a different story. Google SynthID lives in the pixels rather than in metadata, so no metadata reader, this one included, can detect it. It survives cropping, resizing, and re-encoding, only Google's own tooling can read it back, and stripping metadata leaves it untouched. Removing a Content Credentials manifest also removes the proof of provenance a genuine photo carries, so keep the original if that history matters to you.