Uploading a photo of someone's face to a stranger's server in order to protect their privacy is a contradiction. RedactCam runs face detection inside your browser and pixelates faces at full resolution, with you reviewing every box before export.
Two face detectors run together — one tuned for close-ups, one for group scenes — and their results are merged.
Toggle any box off (say, your own face), resize boxes, or drag to add someone the detector missed.
Faces are pixelated into the pixels themselves, and the re-encoded file carries no EXIF, GPS, or camera serial.
School events and kids' sports photos. Protest and street photography. Real-estate shots with neighbors in frame. Marketplace listings. Content for social media where bystanders never agreed to be posted. In each case the person in the frame has a face, and you have an obligation — sometimes legal (GDPR), always ethical — not to broadcast it.
RedactCam is deliberately tuned to over-detect rather than miss: a false box costs you one click to remove, while a missed face published online can't be unpublished. For crowd scenes, review at zoom (mouse wheel) and use a quick drag for anyone the detector didn't catch.
Yes — two detectors run together, one specialized for scenes with many small faces. For very distant crowds, zoom in to review and add any stragglers by dragging.
Detection is strongest on frontal and near-frontal faces. Profiles and occluded faces are detected less reliably, which is why the review step exists — anything missed is one drag away.
We never receive them. Processing happens in your browser's memory; there is no server, no upload, and no account. The app works in airplane mode.
Yes — faces are detected and tracked across frames, pixelated, and the clip is re-encoded in your browser with audio preserved.