Browser-only · nothing uploaded
EXIF, GPS, XMP, IPTC, C2PA and the platform tracking IDs baked into your images, video and PDFs. We actually read your file, show you what's inside, remove it, then re-check. First 3 cleans free.
Loading engine…
The signal platforms read
When you export from ChatGPT, Grok, Midjourney or Firefly, the file carries C2PA Content Credentials and an IPTC/XMP DigitalSourceType tag that says “trainedAlgorithmicMedia.” Meta and others read those tags and attach the label. Barefile finds them and strips them — along with every other hidden field.
Honest caveat: invisible pixel watermarks (like Google's SynthID) are not metadata. No metadata tool — ours included — can remove them. We tell you the truth about what we can and can't do.
Exactly where a photo was taken — street-level.
Camera make, model and sometimes a unit serial number.
Your name, editing app and version, timestamps.
Facebook/TikTok partner IDs embedded in exported video.
Provenance chain some tools sign into the file.
The IPTC/XMP flag that triggers the “AI info” label.
Most “cleaners” just claim the file is clean. Barefile re-reads the output in front of you and shows there's nothing left.
One plan. No credits.
Unlimited cleaning — images, video and PDF. No per-file credits, no packs to top up. 7-day free trial. Cancel anytime.
Card required to start the trial. First 3 cleans are free with no account.
No. Everything happens in your browser. The file never leaves your device — that's why we can keep the privacy promise literally.
We remove AI metadata tags (C2PA, IPTC/XMP source tags) that platforms read for labels. We cannot remove invisible pixel watermarks like SynthID — no metadata tool can, and we won't pretend otherwise.
There isn't one. Because cleaning runs in your browser, each file costs us essentially nothing — so we don't meter it. $20/mo, clean as many files as you want.
JPEG, PNG, WebP and TIFF images, MP4 video, and PDF. More coming.