Blur faces in photos

For photos of school events, street scenes, workplaces or customers that need to go online without showing who was there. A small face-detection model finds the faces on your own device, including the small ones at the back of a group shot, and you check the result before saving.

Loading the tool…

How it works

  1. Drop one photo or many (JPG, PNG, WebP or HEIC). The first time, the face model (0.23 MB) and the runtime it needs are downloaded once.
  2. Check the boxes on each photo: click a box to leave that face visible, or drag across the photo to add a face that was missed. Choose blur, pixelate or a solid colour, and how strong.
  3. Download the photo, or all of them as a ZIP. The copies keep the original resolution and lose the metadata, GPS location included.

Why nothing is uploaded

Every operation on this page is done by code running inside your browser tab, using the same engine that renders web pages. The file is read from disk into your tab’s memory, transformed there, and written back out as a download. It is never sent anywhere — not to us, not to a third party.

Verify it yourself

  1. Open your browser’s developer tools (F12) and select the Network tab.
  2. Load your file and run the tool.
  3. The only requests you will see fetch the tool’s own code — and, for a few heavy tools, their open-source engine (also from this site) or an AI model from Hugging Face — plus one small page-view ping to loreatec.jp (page address and title, nothing more). None of them carry your file.

Proof it stays local →

Frequently asked questions

Are my photos uploaded?

No. The photos are opened, searched for faces and saved inside this browser tab. The only downloads are the face model, from Hugging Face, and the ONNX Runtime engine, from this site — once, and then they are kept for offline use. You can switch the network off after the page has loaded and it keeps working.

Will it find every face?

Not always, and no detector does. It looks at the whole photo and also at overlapping tiles at higher resolution, so the small faces at the back of a group photo are usually found. What it misses most: side profiles and faces turned away, faces partly covered by masks, hands or hair, and very small or motion-blurred faces. Every box is drawn on the preview, so check each photo and draw a box around anything it missed before you publish.

Is the blur strong enough?

At the default settings, yes for the human eye: the blur and the pixel blocks are sized to each face, so a big face in a portrait gets as much as a small one in a crowd, and eyes, nose and mouth do not survive as features. But blur and pixelation keep the shape and colours of a head, and research has shown that software comparing against photos of known people can sometimes still match them. When identification must be impossible — a witness, a child in care — choose the solid colour.

What happens to the EXIF data and the GPS location?

They are removed. The saved copy is a new JPEG (or PNG, for PNG files) written from the pixels, so the camera model, date, GPS location and other metadata of the original are not carried over. That also drops the small preview image many cameras embed in the EXIF data, which would otherwise still show the faces unblurred.

Which model does it use?

YuNet, a tiny face detector by Shiqi Yu’s team, published in the OpenCV Zoo under the MIT licence. The 0.23 MB file is fetched from Hugging Face at a fixed version and checked against its SHA-256 fingerprint before use. It only returns boxes around faces; it cannot tell who anyone is.