Detect and cover faces, plates and private details across photos and video on iPhone, iPad, Mac and Vision Pro - with detection running on the device, so nothing you are trying to protect is uploaded to be protected.
The app runs a detection model on your device to find faces in the image or video, then applies the censor style you chose to each one. Nothing is uploaded to do this - which matters, because uploading photographs in order to hide the people in them would defeat the purpose.
Strong blur, mosaic at a coarse setting, a black bar, or an emoji on a flattened export. Light blur and fine pixelation are not enough on their own: blur is a reversible operation, and small pixel blocks leak the average colour of what was underneath. The practical test is whether you can still tell which way the person is facing.
A lightly blurred one, sometimes. Deconvolution can bring back a recognisable face from a mild blur, and that is ordinary software rather than a research technique. Turn the strength up until the region is a featureless blob and this stops being possible, because the information is genuinely gone.
Use the manual brush to cover it. Profiles, partial faces at the edge of the frame, faces in hats or masks, and faces in reflections are the usual misses. Always review each photo after the automatic pass rather than trusting the count.
Yes. Detection gives you each face separately, so you can keep your own family visible and cover everyone else - which is the common case for a school event or a photo taken in public.
That is what the manual brush is for. Worth scanning every photo for: plates on cars and in reflections, house numbers and street signs, laptop and phone screens that are readable at full size, and anything with a badge or barcode.
Yes. Faces are tracked across frames rather than being detected once, which is the part that makes video practical at all. Check the result at the points where someone turns away or leaves the frame - tracking is at its weakest exactly there.
This is worth handling deliberately. A photograph carries GPS coordinates and a timestamp in its EXIF data, and those survive cropping and blurring - so a picture taken in your garden can publish your address even with every face covered. Most social platforms strip EXIF on upload; most messaging apps do not when sending as a file, and email attachments certainly do not.
No. Redaction always produces a new file. Keep it that way as a habit: once you have edited over the only copy, the decision about whether to share the unblurred version has been made for you permanently.
On a flattened JPEG or PNG, yes - the pixels underneath are gone. The risk is exporting to a format that keeps layers, or sharing the original alongside the edit by accident, in which case the sticker is an object somebody can move aside. Export a flattened copy and the question disappears.
No. Detection, tracking and rendering all happen locally on iPhone, iPad, Mac or Vision Pro. There is no account and no upload, and the app works with no network connection.
iOS 17 or iPadOS 17 on iPhone and iPad, macOS 14 or later on Mac, and visionOS 1.0 or later on Apple Vision Pro. One purchase covers every platform.
Automatic face detection plus a manual brush for plates, IDs and screens - blur, pixelate, mosaic, emoji or black bar, on photos and video.
The order that catches the things people miss, which are rarely the faces.
If you are publishing a group photo and do not know who has consented, cover everyone who is not yours. It costs nothing and removes the question entirely.
Try this in Face Blur AI - Hide & Censor →Tracking does the work, and the review is where the mistakes are caught.
A single uncovered frame is enough. Pausing at random points in the export is a faster check than watching the whole clip.
Try this in Face Blur AI - Hide & Censor →Our support team usually replies within one business day.
Download Face Blur AI - Hide & Censor Contact support