Remove Hidden Metadata From WebP Images
04 Google's format, increasingly everywhere
WebP shows up constantly now - Chrome, Android, and a growing share of websites use it by default because it's smaller than JPEG or PNG at similar quality. That means most WebP files people encounter were downloaded, not shot on a camera - the metadata question is less "did my phone leak my GPS" and more "does this image I saved carry something from wherever it came from." WebP is a RIFF container (the same family as .wav and .avi), with EXIF, XMP, and ICC profile data each stored as their own named chunk.
Go deeper: RIFF chunk structure, and the one honest limitation here
Like every other container format on this site, WebP's RIFF header declares its own total size - and if the actual file is larger than that declared size, that's exactly the appended-data pattern covered on the comparison page, checked and reported the same way.
The one real, disclosed limitation: this tool's WebP encoder is a pure-Rust, lossless-only implementation (deliberately - it avoids a C dependency the rest of this pipeline doesn't need elsewhere). Animated WebP input gets flattened to its first frame on output as a result. Static WebP - the overwhelming majority of what you'll actually encounter - is unaffected either way.
Why this still matters for a "downloaded" image
Even an image you didn't take yourself can carry metadata worth knowing about - a content-provenance manifest (C2PA / "Content Credentials") noting it was AI-generated or edited, an ICC profile identifying the software that produced it, or EXIF fields nobody bothered to strip before uploading it somewhere you later saved it from. This tool detects and reports a C2PA manifest specifically before removing it, the same as it does for JPEG and PNG - covered in more depth on the AI training page.