Clean the metadata off a batch of files
Select a folder's worth of photos and audio at once. Each file is read, its metadata stripped and reassembled individually — you get a report of what came off what, and the cleaned files download together.
The leak is rarely one file
If a single photo tells someone where you live, a folder tells them where you have been. Location, timestamps and device serial numbers combine into a pattern, and the pattern is the part worth removing — which means doing all of it, not the one photo you noticed.
Typical cases: photos before posting a listing, a set of scans before sending to a client, voice memos before sharing a project folder, or a whole camera roll before handing over a drive.
Per file type
| Format | Removed | Kept |
|---|---|---|
| JPEG | EXIF (camera, lens, serial, GPS, timestamps), XMP, IPTC / Photoshop block, C2PA / JUMBF, comment segments | Image data, JFIF header, ICC colour profile, the rotation flag |
| PNG | eXIf, tEXt, iTXt, zTXt,
caBX, tIME |
All image chunks, palette, transparency, colour profile |
| WebP | EXIF and XMP RIFF chunks |
VP8 / VP8L / VP8X image chunks — with the container size rewritten correctly |
| MP3 | ID3v2 (including embedded cover art), ID3v1, APE tags | All audio frames, byte for byte |
Files that fail to parse are listed with the error rather than silently skipped — you should know which ones did not go through.
Things worth knowing before a big run
- Your originals are never modified. This reads the file you selected and produces a new one. Nothing is written back to disk in place.
- Downloads may ask permission. Browsers block rapid consecutive downloads, so a large batch will prompt. Allow it once and the rest go through.
- Very large batches take a moment. Files are processed one at a time so memory stays flat and you can see progress.
- Nothing is uploaded. No file leaves the browser — you can watch the network panel to confirm.