Blur Face in Photo – Automatic, Free and Private

Add a photo above and the faces in it are found and marked for you in a few seconds. Choose how to cover them: Blur, Pixelate, Solid box or Emoji. Check the marks, add any face the tool missed by dragging over it, and save. It is the quickest way to blur face in photo files before you post them, send them or put them in a report.

Everything happens in your browser. The photo is opened, searched for faces, covered and saved on your own device, and it is never uploaded. JPG, PNG, WebP and HEIC photos from iPhones work, and most browsers also open GIF, BMP and AVIF. The saved picture keeps the full resolution of the original.

To blur face automatically, the tool uses YuNet, a small face finding model from the OpenCV Zoo collection, released under the MIT licence. The model file is only about 230 KB. It is downloaded once from our site into your browser and run there with ONNX Runtime, so your photo never leaves your device.

YuNet looks at the picture at 640 × 640 pixels. A big photo is first checked as a whole, then again in overlapping tiles at a higher resolution, so small faces at the back of a group are found too: down to about one fortieth of the photo's width. Large photos are searched on a copy of at most 4000 pixels, and the marks are then placed on the full-size picture.

It finds most faces seen from the front and from the side. Faces that are very small, turned away, partly hidden by a mask or a hand, or heavily blurred can be missed, and now and then a pattern or a poster gets a mark it should not have. That is normal for any automatic finder, so always look over every photo before you save it. Your eyes are the final check.

Open More options and the card Find faces automatically has two settings. When photos are added is ticked by default; untick it if you prefer to mark every area by hand. Sensitivity goes from 20% to 90% and starts at 60%.

A higher sensitivity finds smaller, turned and blurry faces too, but it also gives more false marks. A lower value gives fewer marks that are almost always real faces. After you change it, press Find faces again above the photo: the automatic marks are replaced, and the ones you drew yourself are kept. Remove all marks clears both kinds.

Every mark on the photo is a button. Green means it was found automatically, blue means you drew it. Click a mark once to skip it: it turns dashed and that area stays as it is in the saved photo. Click it again to cover it again. Double-click a mark to delete it completely.

To add a mark, press and drag on the photo. A red dashed frame shows the area while you drag; let go and it becomes a blue mark. This works for anything, not just faces: number plates, name badges, house numbers, a computer screen or a letter on the table. On a touch screen, drag with one finger.

The small number on each thumbnail shows how many areas will be covered in that photo, and the line above the preview says how many faces were found and how many you marked by hand. If nothing is marked, the button note reminds you before you save.

A few habits make checking quicker. Look at the edges of the photo first, where people are often cut in half and harder to find. Look for faces in reflections: mirrors, windows, sunglasses and phone screens can all show a face that the finder treats as part of the background. Photos and posters on a wall, faces printed on T-shirts and people on a television in the background may need a mark too, or may get one you do not want.

Both methods hide detail, but they look different and behave a little differently. If you pixelate face areas, each block takes the average colour of the pixels under it, which gives the familiar TV mosaic. Blur spreads every pixel over its neighbours, which looks softer and more natural in a photo.

For privacy, the size of the effect matters more than the type. Large pixel blocks or a strong blur leave too little detail to read a face. Small blocks and a light blur keep shapes that people, and software, can sometimes still recognise, especially for someone they know. Keep the strength at 60% or more when identity matters; the hint in the tool says the same.

For a whole picture rather than single faces, Pixelate Image and Blur Image have more control over the full frame.

A blur face emoji cover is the friendly option: a smiley, a pair of sunglasses, a cat or a heart sits where the face was. Parents use it for children in birthday photos, schools and clubs for pupils whose families did not agree to photos, and friends for anyone who would rather not be online.

Choose Emoji under Cover with and tap one of the eight choices. Each emoji is drawn at the size of its mark, plus the extra margin, so a bigger mark gives a bigger emoji. Emoji look the way your device draws them, so the same smiley can look a little different on iPhone, Android and Windows.

Like a solid box, an emoji covers the face completely, so it is one of the safest choices. Shape, Soft edge and Blur everything except the marks do not change emoji; they are made for the other three methods.

Oval is the default because faces are roughly oval, so the cover hugs the head and leaves less of the background touched. Rectangle covers the corners of the mark as well, which is better for number plates, documents and screens, and is what the two box presets use.

Soft edge feathers the border of Blur and Pixelate covers so they fade into the photo instead of ending in a hard line. It is on by default. A solid box always has a hard edge, so nothing of the face shows through.

Extra margin around each face goes from 0% to 60% and starts at 15%. Automatic marks fit tightly around the face, so a little margin also covers hair, ears and the outline of the head, which can give a person away as much as the face itself. Raise it for hair that stands out; lower it when faces are close together.

Tick Blur everything except the marks to turn the effect around: the marked areas stay sharp and the rest of the photo is covered. With Blur you get a portrait look with a soft background; with Pixelate a mosaic background; with Solid box a plain colour all around.

It is useful when you want to keep one person and hide everyone else. Click the marks of the people who should be hidden to skip them, leave the one person you want to keep covered, and with this option on, only that person stays visible. It also works for a product, a sign or any area you mark by hand.

Group photos are where an automatic finder saves the most time. To blur faces in group photo shots from a class, a team, a party or a crowd, add the photo and let the tiled search mark the rows of faces, then zoom in with your eyes and check the back rows, people in profile and faces half hidden behind others.

Crowds often contain people who look down or away. Raise the Sensitivity to 75% or 80% and press Find faces again, then delete any false marks with a double-click. If faces are packed tightly, lower the Extra margin so the covers do not merge into one patch, or switch to Rectangle.

To keep only one person, for example your own child in a school photo, use Blur everything except the marks as explained above, or skip that one mark and cover all the others.

Most tools that blur face online send your photo to a server, where it is stored for some time. Here the page loads once, and from then on the photo is read, searched and covered by your browser on your own device. No file, no preview and no copy is sent to us.

That matters for this job in particular, because the photos that need faces hidden are often the private ones: children, patients, witnesses, colleagues or strangers in the street. Your last settings are kept in your browser's own storage, and you can clear them at any time.

You can blur face free here with no account, no email address and no app to install. There is no watermark on the saved photo and no daily limit.

It costs nothing to run because your own device does the work. Only the page and the small model file come from our site; finding and covering faces happens on your phone or computer, even when the connection is slow.

Select many photos at once or drop a pile on the box: up to 50 photos go into one batch. Each photo is searched for faces as it opens, and the thumbnails above the preview let you step through them one by one to check the marks. The cover settings are the same for all photos; the marks are separate for each.

Press the button, which reads Blur faces or, for a batch, Blur 12 photos, and every photo is saved at full resolution. Download all (ZIP) saves them in one archive named blurred-photos.zip; with a single photo the button is Download photo. Each file gets -blurred added to its name. Change marks clears the results so you can adjust and save again without adding the photos a second time.

Under More options, Format starts at Same as original. PNG, GIF and BMP photos are saved as PNG, WebP as WebP, and JPG and HEIC as JPG. You can also choose JPG, PNG or WebP for all photos; JPG and WebP are saved at high quality.

The saved photo is drawn fresh from the covered picture, so it no longer contains the camera details and location (EXIF) of the original. That is a second layer of privacy: the GPS position in a phone photo can reveal a home or a school even when every face is hidden. To clean other photos without covering anything, use Remove Metadata.

Which format should you choose? Keep Same as original unless you have a reason to change it. JPG gives small files that every site and app accepts, which suits photos for social media and messages. PNG keeps every pixel exactly and is the better choice for screenshots with text. WebP gives small files for websites. Whatever you pick, the photo keeps its full pixel size.

To blur face on iPhone, open this page in Safari or any other browser, tap Select photos and pick from Photos or Files. HEIC photos work directly. Faces are marked automatically; drag with one finger to add a mark and tap a mark to skip it. The saved photo goes to Downloads in the Files app, and Share, then Save Image, moves it into Photos.

On Android, pick from the gallery or a file manager; saved photos land in the Downloads folder. On Windows and Mac, drag photos from a folder or paste a copied picture with Ctrl+V. Face finding runs on the processor, so a phone may take a few seconds longer per photo than a computer, and for a batch of large photos a computer is more comfortable.

If the goal is to hide face in photo files so that nobody can tell who it is, choose a cover that removes the face completely. A Solid box or an Emoji leaves nothing behind. A light blur or small pixel blocks can sometimes be recognised, especially by someone who knows the person, and the outline of the hair, clothes, tattoos or the setting can be enough to identify someone.

So for real anonymity: use a solid box or an emoji, or blur and pixelate at a high strength; add extra margin so hair and ears are covered; cover name badges, uniforms with names, number plates and house numbers by hand; and share the saved copy, never the original. Keep the original somewhere safe or delete it, as you need.

Think about what else is in the frame as well. A school logo on a jumper, a street sign, a car number plate or the view from a window can tell people where a photo was taken and who is in it. Drag a mark over each of these. The same cover method is used, so a solid box over a face and a solid box over a sign look the same and draw no extra attention.

Many people want to anonymize faces in photos before sharing them because the law or their workplace asks for it. In the European Union and the UK, a recognisable face is personal data under the GDPR, and publishing it can need a legal basis such as consent. Many other countries have similar privacy or image rights, and schools, hospitals and employers often have their own photo rules.

Covering faces is a good habit for photos of children, patients, people at protests, witnesses and bystanders who did not agree to be shown. This page gives general information, not legal advice; if you are unsure what applies to you, check the rules of your country or ask your organisation.

This tool works on still photos only; video is not supported. To hide faces in a short clip, one way is to turn it into frames or a GIF first. For documents, Redact PDF blacks out text and areas in a PDF, which is the better tool for scanned forms and contracts.

Each photo is opened by the browser and turned the right way up. A copy is fed to YuNet in ONNX Runtime, as a whole and in tiles, and overlapping results are merged into one mark per face. When you save, the photo is drawn at full size, a covered layer is made with blur, pixel blocks or a colour, and a mask with your ovals or rectangles cuts it to shape, with a feathered border when Soft edge is on. Emoji are drawn on top as text. The result is saved as JPG, PNG or WebP.

After you hide the faces, Crop Image trims the photo to the part you want to share, and Compress Image makes it smaller for upload limits. Remove Metadata cleans photos you did not need to cover, and Pixelate Image and Blur Image work on the whole picture.