Blur Faces in Photos

Anonymise a photo before you share it: drop it in and every face it finds is blurred, pixelated or blacked out, right in your browser. Click a face to leave it visible, drag a box over any face it missed, and save the result or a ZIP of a whole batch. Your photos never leave your device.

Effects

EffectLookAnonymity
BlurSoft, naturalGood at high strength
PixelateMosaicGood with large blocks
Solid blockBlack oval or boxComplete

When to use it

  1. Hide bystanders' faces before posting street or event photos.
  2. Protect children's identities in school or family pictures.
  3. Anonymise people in protest or news photos.
  4. Redact faces in screenshots or documents for reports.

How to Use

  1. Drop one or more photos; faces are found automatically.
  2. Choose blur, pixelate or solid, and the strength and margin.
  3. Click faces to keep them visible, or drag to cover missed ones, then save.

Features

  1. Automatic face detection
  2. Tiled detection for small faces
  3. Blur, pixelate or solid block
  4. Oval or rectangle
  5. Toggle each face
  6. Draw your own boxes
  7. Batch with ZIP
  8. Never uploaded

FAQs

Does it find every face?

Not always. Side profiles, faces in shadow and very small faces can be missed, so always check. You can drag a box over any face it missed.

Can I keep some faces visible?

Yes. Click a green box to leave that face visible.

Which effect is safest?

A solid block. Strong blur or large pixels also remove the detail, but light blur can leave a face recognisable.

Are my photos uploaded?

No. Detection and blurring run in your browser. Only the detector program and model are downloaded, never your photo.

Can I do many photos?

Yes. Drop several, check each, and download them all as a ZIP.

Technical Details

Faces are detected with Google's MediaPipe BlazeFace short-range model, running on WebAssembly in your browser; the photo is never uploaded. BlazeFace looks at a 128×128 view of the image, so a small face in a large group photo can be too small for it to see. The tool therefore runs it on the whole photo and again on overlapping 2×2 and 3×3 tiles. Detections of the same face from different tiles are merged into one box covering all of them. Each box is widened by an adjustable margin to cover hair, ears and chin. Blur and pixelate both shrink the face to a few blocks and scale it back up (smoothly for blur, sharply for pixelate), which removes the detail rather than hiding it. A solid block replaces it entirely. The confidence slider filters the stored detections without re-running the model. The output is the full-resolution photo; PNG and WebP keep their format, other photos save as JPEG at quality 92.

Image Tools

Image Tools

  • Watermark Remover
  • Screenshot Annotator
  • Image Metadata Remover
  • JPG to Text
  • PNG to Text
  • Screenshot to Text
  • SVG Optimizer
  • HEIC to JPG