Blur Faces in a Photo
Blur Faces finds the faces in a photo with an AI model running in your browser and blurs, pixelates or blocks them out - toggle any box or add ones it missed. The photo is never uploaded.
Paste image (Ctrl+V), drag and drop, or click to upload
Image file · processed on your device, never uploaded
Faces are found on your device. The first time, your browser downloads the face detector (about 4.6 MB).
At a glance
- Accepts
- PNG, JPG, WebP, GIF and other images your browser decodes
- Outputs
- The same image with faces blurred
- Where it runs
- Entirely in your browser, on your device
- What leaves your device
- Nothing - the file never leaves your device
- Limits
- No size or count cap; bounded by your device's memory
- Account
- Not required - there is no sign-up
- Price
- Free, with no watermark and no usage cap
- Made by
- Visrut
- Last updated
- 12 October 2026
Hide every face in a photo before you post it - a class photo, a protest, a street scene, a screenshot of a group chat. An AI face detector (Google's MediaPipe BlazeFace) finds the faces on your own device, then each one is blurred, pixelated or covered with a solid block. The photo is never uploaded. To catch small faces in group shots, the detector scans the whole picture and then overlapping close-ups of it, and anything it isn't sure about is offered as a dashed box you can switch on.
How to use it
- Paste (Ctrl+V), drag and drop, or click to upload a photo.
- Wait a moment while faces are found - the first run downloads the detector, about 4.6 MB.
- Click any box to turn it on or off. Use Add a missed face to drag a box over a face the detector didn't catch.
- Pick blur, pixelate or solid block, and adjust strength and padding.
- Check the result, then download it or copy it to the clipboard.
Limits
- BlazeFace is built for faces looking roughly towards the camera. Side profiles, faces turned away and faces in deep shadow are often missed - always look over the result before sharing.
- Faces smaller than about 1/40 of the photo's shorter side may not be found. Add those by hand.
- Still images only. For clips, use Blur Faces in Video.
Frequently asked questions
Is my photo uploaded to find the faces?
No. Face detection runs on your own device with Google's MediaPipe BlazeFace model, inside your browser. The first time you use the tool your browser downloads the detector (about 4.6 MB: the MediaPipe runtime from the jsDelivr CDN (or unpkg if jsDelivr can't be reached) and the model file from Google's model storage), then caches it. Your photo itself never leaves the page.
Can someone un-blur the faces afterwards?
Not from the downloaded image. Every effect first shrinks each face down to between 4 and 16 pixels across and then scales it back up, so the detail is thrown away rather than just softened. A light Gaussian blur can sometimes be partly reversed - this can't, because the information is no longer in the file. Solid block replaces the face with a flat colour.
It missed a face. What can I do?
Turn on Add a missed face and drag a box over it. Faces the detector was unsure about are already shown as dashed boxes - click one to blur it too. Side profiles, faces turned away, heavily shadowed faces and very small faces in large crowds are the ones most often missed, so check the result before you share it.
It blurred something that isn't a face.
Click the box to turn it off. The area goes back to the original pixels.
Which is safest - blur, pixelate or solid block?
Solid block hides the most, since nothing of the face is left. Blur and pixelate at the default strength leave no recognisable features; push the strength slider up for fewer, larger cells. Increase padding if hair or ears should be covered too.
What image formats work?
Anything your browser can open - JPEG, PNG, WebP, GIF and more. The download keeps the original format. HEIC photos from an iPhone need converting first in most browsers other than Safari.
Does it work on video?
This page is for photos. For video, use Blur Faces in Video (/blur-faces-video), which finds and follows faces through a clip and exports an MP4, also without uploading it.