Drop a photo and we'll detect faces and pixelate them automatically. Add or remove blur regions by hand. Helpful for blogs, news posts, public Slack channels — anywhere you want to share without doxxing anyone.
JPG, PNG, WebP, HEIC. We run a face-detection model locally to find faces.
Add boxes for faces we missed (e.g., side profiles). Remove boxes for false positives. Adjust blur strength: pixelated, gaussian, or solid.
All other pixels are identical to the original. Only the blur boxes are altered.
Single photos: this tool. Whole camera roll? SwipePhotos works directly on your Apple Photos library — no exporting and re-importing. Swipe through years of bursts and duplicates in a weekend. 100% on-device.
Entirely in your browser. We use MediaPipe (or face-api.js) compiled to WASM. Photos never leave the page — important if you're redacting something sensitive like a medical photo or a court document.
Mostly. Front-facing faces are caught with high accuracy. Profile shots and partially obscured faces are harder; that's what manual boxes are for. Always review every output before publishing.
It depends on the pixel size. Small pixelation (12×12 px or finer) has been reversed by AI in research papers. Use the 'Solid black' or 'Heavy blur' option for true anonymity. Pixelation is fine for casual privacy on social posts.
Yes. Switch to manual mode and draw boxes over anything you want hidden. Useful for license plates, screen contents, name tags, addresses on packages.
Only inside the blur boxes. The rest of the image is pixel-identical to the original.