A face cover fails in two quiet ways. The first is strength: a polite smear that still shows eye spacing, a jawline, or the outline of glasses is still a recognisable person to anyone who knows them. The second is context: a jersey number, a name badge, a door plate, a tattoo, or a reflection that reprints the same identity after the face is washed out.
This page is the still-photo face walkthrough for BlurThis. You open the blur a face tool, run Find faces or place Box and Draw yourself, push cover until the region is an even wash or a deliberate mosaic at full zoom, then hunt those second identifiers. Face blur is not identity-proof. A heavy cover flattened into a new PNG is a strong casual barrier. It is not a guarantee against someone who already knows the person, and it is not a forensic wipe of device memory or cloud backups.
Blur face in photo online free (strength first)
Use Find faces for suggestions, then Blur or Pixelate hard enough that the face is unreadable at 100% zoom. Use Draw for hair, hats, and turned heads. Use Black bar for name tags and jersey digits.
In BlurThis
These steps are about face strength and context leaks on a still. Open the blur a face tool, then:
- Choose the photo. Tap Find faces and accept the boxes you mean to cover, or drag Box / Draw yourself for profiles and misses.
- Set Blur or Pixelate until the face is an even wash or large tiles at full zoom, not a soft haze that still shows features.
- Bar name tags, jersey numbers, and ID strings with Black bar. Check reflections, door plates, and bags for a second copy of the person.
- Tap Download PNG. Open the export at full size, then send that file, not the original in Recents.
Open larger image in a new tabStrength: how hard is hard enough?
Stronger than looks polite, and still not identity-proof. Shape of the hairline, the jaw, glasses, a beard, and skin tone survive a gentle smear. Push until the region is an even wash of colour, or switch to pixelate if you want the cover to look deliberate. Zoom until an ear or glasses frame would be obvious if it leaked. If you can still name the person from outline alone, raise the strength or enlarge the region.
Blur and pixelate are different looks, not a ranked upgrade at the same slider. Soft blur blends into a group photo. Pixelate reads as censored on purpose, which is often what you want in a news-style or documentary crop. Both fail when set too light. The side-by-side comparison lives on pixelate or blur faces in a picture; this page owns the decision to cover a person before you post.
Open larger image in a new tabFind faces, then finish with your eye
Find faces speeds crowded group shots because it suggests boxes in one pass. The detector runs in the page. It may download a small model to you the first time; a wait is local work, not proof that the PNG left the device. Treat suggestions as a first pass. A profile, a faint reflection, a face half behind a hat, or someone far in the background may still need your own mark.
Accept the strangers you mean to cover. Leave the people who consented. For a child, cover the face and immediately check for a name on a jersey, a bag, or a classroom board behind them. Group photos, school events, parties, protest crowds, and a classroom in the background of a work shot all share the same problem: not everyone in the frame agreed to be posted. Detector boxes help. Your eye finishes the job.
For a turned head or a messy hairline, switch to Draw and close a freeform loop so the cover follows the silhouette instead of leaving a sharp rectangle of skin. Box remains the fast path for a front-facing portrait sitting in a clean rectangle of the frame.
Context leaks that outlive a washed face
Covering the face without a second pass is unfinished work. Walk the rest of the frame for jersey numbers, tattoos, door numbers, plates, distinctive jackets, lanyards, and window reflections. Any of those can identify someone as surely as a sharp nose. Treat each as its own region. Use Black bar on short text. Digits and letters survive soft blur at full zoom in a way a face often does not, because the alphabet is small and the font is known.
If the photo is a screenshot of a video call, the nameplate under each tile is the badge. That layout has its own guide: blur names on a Zoom or Meet screenshot. If you only need one tight region on a general screenshot, the partial cover skill is how to blur part of a screenshot.
Hiding yourself, or cropping out entirely
The common case is not a stranger but you: a group photo you want to post without being in it, a background cameo in someone else's shot, a marketplace selfie where the item matters and your face does not. The tool steps match. The judgement differs. Ask what else in the frame identifies you (jacket, tattoo, front door, street sign), then cover those regions as separate boxes.
Cropping is often cleaner. If you sit at the edge of the group, cutting yourself out removes pixels that no blur has to defend later. Pixels that never enter the export cannot be reconstructed from that file.
Phone browser path without an install
Safari on iPhone and Chrome on Android run the same editor as a desktop. There is no App Store or Play Store download and no account. Portrait mode on the camera only softens the background when you take the shot. Photos Markup has no blur brush; a filled opaque shape can cover a badge if you prefer to stay in Markup, but soft highlighters are not a face cover. Full device walkthroughs live on the iPhone and Android guides linked from the how to blur hub; this page stays on the strength and context checklist.
When face cover is not enough
Name tags, jersey numbers, licence plates, and any readable string in the same frame need a black bar or a crop. If the image came off a camera rather than a screenshot, redacting pixels does not strip GPS from EXIF by itself; export a fresh PNG from the editor or strip metadata in the share sheet. A flattened PNG with replaced pixels is the file you share. Later JPEG compression cannot reconstruct the original face from those replaced pixels. Weak blur, a missed edge, or a name left in the frame is a separate problem.
Covering a photo is not HIPAA or GDPR compliance by itself, and it is not a forensic wipe of disk, backups, or chat history. Closing the tab drops the working copy from this page only.
Faces that move (video)
On a clip the face travels across many frames. You place timed boxes with Blur In and Blur Out, chain segments when the head moves, and cut high-risk seconds when cover is the wrong answer. Video has no opaque black bar. Limits include 50 MB input, 90 seconds export, and 720 pixels on the longest edge. Do not paste that whole recipe here; open blur a face in a video online free when the file is a recording rather than a still.
FAQ
Can someone un-blur a face? Tools that claim to “unblur” a face generate a plausible face, not a guaranteed restore of the original. A light blur is the easy case for recognition by someone who knows the person. A heavy cover that fully replaces the pixels, then flattened into a new file, does not leave those original pixels sitting under a filter. Hair, clothes, and a name tag can still identify them. Face blur or pixelate is not identity-proof.
Does Find faces upload the photo? No. The detector is a small model that runs inside the page. The first time, your browser may download the model file; after that, detection can work with the network off. Offline use after the page has loaded shows local capability. It does not prove that an online session never loads other assets. On a computer, watch the Network panel, including images and media, not only Fetch/XHR.
Watermark or limit? No watermark, no account, no cap on how many photos you run. Images export as PNG. Prefer sending the download. Hub: how to blur.
