Skip to content
TabShedKitsMy desk

How to Blur Faces Before You Share a Group Photo

Updated 5 min readBy

A privacy checklist for hiding faces in photos: which style to choose, how strong to blur, and how to catch the faces the detector misses.

Blur FacesFind and blur faces in photos
Open Blur Faces
On this page
  1. What the tool does and does not do
  2. Workflow, step by step
  3. Choosing a style
  4. Padding and shape
  5. A final check that catches most mistakes
  6. When this is not enough
  7. After blurring
  8. A worked example: a classroom photo
  9. FAQ

You want to post a picture from a school event, a street scene or a team outing, but not everyone in it agreed to appear online. Hiding faces sounds simple until you look at the result: one face was missed in the back row, a light blur still looks recognisable, and the photo now looks messy. This guide walks through a careful workflow in the Blur Faces in Photos tool, with the checks that matter most when privacy is the goal.

What the tool does and does not do

The tool runs a small face detector inside your browser. The first time you use it, the detector files (about 10 MB) are downloaded once and then kept by the browser, so later runs also work offline. Your photos are not uploaded. Big pictures are scanned in overlapping tiles so small faces in group shots are found as well.

The detector is not perfect. It will usually find faces that look toward the camera, are lit and are larger than a few pixels. It can miss very small faces (under roughly 15 pixels), faces turned away, faces behind a mask or in deep shadow. Because of this, treat the automatic step as a first pass and always finish with a manual check.

Workflow, step by step

  1. Add the photos. Drop one picture or a whole batch. Up to 40 photos are accepted at a time, as JPG, PNG or WebP; extra files are skipped and a message tells you so.
  2. Wait for detection. Each photo shows how many faces were found, or that none were found.
  3. Review every face. Faces appear as frames on the photo. Click a frame to turn that face off (leave it visible) or on again. If you want only the strangers blurred and your friends clear, this is how you do it. The Blur all and Unblur all buttons switch everything at once.
  4. Add missed faces. Drag a box on the photo over any face the detector skipped. Manual boxes are listed as Box 1, Box 2 and so on, and are treated like the automatic ones.
  5. Pick a style and strength (see below).
  6. Download. A single photo downloads on its own; several come as one ZIP. The faces are changed in the saved pixels, they are not hidden behind a layer that someone could remove.

Choosing a style

StyleLooks likeBest when
BlurSoft smear, with a Blur strength sliderYou want a natural look and the face is not the focus
PixelateCoarse blocks, with a Pixel size sliderA familiar "censored" look; blocks stay obviously hidden
BlackSolid boxMaximum privacy, for example for children or sensitive scenes
EmojiA sticker such as a smiley or a hidden-face emojiFriendly social posts where a solid box would feel heavy

The strength slider runs from 5% to 100%. A low blur on a large face can still be recognisable to someone who knows the person, because the overall shape, hair and skin tone remain. If it matters, move the strength up until you cannot tell who it is, or switch to the Black or Emoji style, which remove the information altogether.

Padding and shape

Detectors draw their box tightly around the face. Hair, ears and chin can stay outside it. The Padding around each face slider (0% to 80%) enlarges the covered area. A bit of padding makes the result look more deliberate and hides head shape. The Shape control switches between Oval, which follows the face more naturally, and Box, which covers the corners as well.

A final check that catches most mistakes

  • Zoom into the corners and edges. People at the edge of a photo, partly cut by the frame, are often missed.
  • Look at reflections: mirrors, windows, glasses and phone screens can show faces too.
  • Check name badges, number plates and screens in the same photo. The tool only handles faces.
  • Look at the photo at normal size on a phone. If you can guess who someone is, strengthen the effect.

When this is not enough

Blurring a face does not remove other identifying details such as a distinctive tattoo, an unusual outfit or the location. Photo files can also carry hidden data; the tool rewrites the picture, but if you share an original elsewhere that detail stays with it. For video clips use the Blur Faces in Video tool. If you only need a whole-photo soft focus, such as a background for text, the Image Blur tool is simpler.

After blurring

JPEG photos are saved again at high quality, while PNG and WebP keep their format. If the pictures are going on the web, a pass through the Image Compressor makes them lighter, and the Image Resizer sets exact dimensions for a platform.

A worked example: a classroom photo

Suppose you want to post a photo of a school trip where only your own child may be shown. Drop the photo, wait for the face count, then press Blur all and click your child's frame to switch it off. Drag a box over the child in the back row that was missed because of a hat brim. Choose Pixelate or Black with a generous padding, check the corners for partial faces, and download. Looking at the finished photo at phone size is the last test: if you can tell who the blurred children are from their clothing, consider cropping the picture tighter instead.

FAQ

Can someone undo the blur?

The downloaded file contains the changed pixels, not the originals. Still, a light blur may be guessable, so use a higher strength or a solid style for anything sensitive.

It found a face that is not a face. What now?

Click the frame to switch it off.

Does it work on a phone?

Yes for a few photos. Large batches of big pictures are faster on a computer.

Does it change the rest of the photo?

Only the face areas change.