Blur faces in photos automatically
Sharing a photo from a school event, a protest, a street scene, a party or a workplace often means showing people who never agreed to be online. This face blur tool finds faces automatically and hides them in one step — with a strong blur, pixelation, a solid patch or a fun emoji — so you can post the picture without exposing anyone.
Detection uses Google’s MediaPipe face detector (BlazeFace), an Apache-2.0 licensed model that runs inside your browser. Tap any box to keep that face visible — your own, say — drag to cover a face it missed, and process a whole batch of photos at once, downloading them together as a ZIP. Photos are never uploaded.
How to use the Face Blur
- Drop one or more photos onto the upload area, or click to choose them. Faces are detected straight away.
- Choose the effect — Blur, Pixelate, Emoji or Solid — and an oval or rectangular shape. Adjust the strength and the padding around each face.
- Check the photo. Tap a box to switch that face off or on, drag on the photo to cover a face that was missed, and press × to remove a box.
- For group photos keep “Also scan for small faces” ticked; switch sensitivity to “Catch more” if faces are missed, or “Strict” if objects are boxed by mistake.
- Press Download this photo, or Download all (ZIP) when you added several photos.
Face Blur features
Automatic face detection
MediaPipe’s BlazeFace short-range model locates faces and their landmarks in milliseconds, on your device.
Small-face scan
Besides the whole photo, overlapping tiles at decreasing sizes are scanned so faces in the back row of a group photo are found too.
Four ways to hide
Strong blur, chunky pixelation, an emoji over each face (with a pixelated layer underneath), or a solid patch.
Full control
Switch individual faces off, add boxes by hand, change the padding, and pick oval or rectangular coverage.
Batch processing
Add several photos, review each one from the thumbnail strip, and download them all in a single ZIP.
Private
Detection and editing happen in your browser; the saved images contain no EXIF data such as GPS location.
The model and its limits
Faces are found with BlazeFace short-range from Google’s MediaPipe Tasks (Apache-2.0 licence for both the library and the model, which is downloaded from Google’s model storage). It is a fast single-shot detector designed for faces roughly within two metres of the camera, and it reads a 128 × 128 pixel version of the image. That is why this tool also scans overlapping tiles of the photo: a face that is tiny in the whole picture becomes large enough to detect within a tile. Faces smaller than about 1/80 of the photo’s shorter side on small photos, or about 1/60 on 12-megapixel photos, can still be missed.
Like every face detector, it struggles with faces in profile or turned away, partly covered by hands, hats, masks or hair, strongly blurred by motion, very dark, or very small. It also does not look at reflections in windows and mirrors, faces on posters or screens, or other identifying details such as name badges, tattoos and number plates. Always review the result and add boxes by hand where needed.
Choosing an effect
| Effect | Best for | Note |
|---|---|---|
| Blur | Natural-looking photos for news, blogs and social media | Use a high strength; light blur can leave a face recognisable |
| Pixelate | A clear “anonymised” look | Larger blocks hide more; strength sets the block size |
| Emoji | Family and party photos, children’s faces | A pixelated layer sits underneath, so the face stays hidden even if the emoji font is missing |
| Solid | Maximum certainty, legal or evidence use | Nothing of the face remains |
Before you post a photo of other people
- Ask for consent where you can, and always for children — many schools and clubs require faces of children to be hidden.
- Check for other identifiers: name tags, school logos, house numbers and number plates. Cover those with Pixelate Image or Redact Image.
- Original phone photos can contain GPS coordinates. The files saved here have no EXIF data; check others with the EXIF Viewer.
- Cropping a bystander out with Crop Image is often better than blurring them.
Face Blur FAQ
How do I blur faces in a photo automatically?
Drop the photo into this face blur tool. Faces are detected and blurred immediately; check the result, tap any face you want to keep visible, and download.
Why was a face not detected?
Small, turned, covered, dark or blurred faces are hard for any detector. Keep “Also scan for small faces” on, try “Catch more” sensitivity, and drag a box over any face that is still missed.
Can I keep one face visible and blur the rest?
Yes. Tap the box around the face you want to keep — it turns dashed and is no longer hidden. Tap again to hide it.
Can I blur faces in several photos at once?
Yes. Add as many photos as you like; each is scanned in turn. Review them from the thumbnail strip and press “Download all (ZIP)”.
Is face blur reversible?
A strong blur, large pixel blocks, an emoji or a solid patch cannot be undone in the saved file. Very light blur can leave a face recognisable, so keep the strength high.
Are my photos uploaded or used for face recognition?
No. Detection only finds where faces are; it does not identify anyone. It runs entirely in your browser and the photos never leave your device.