Censor Image and Pixelator

Hiding sensitive details – faces, license plates, ID numbers, credit cards, confidential document fields – while leaving the rest of the image intact. The mosaic effect makes the covered region visually unreadable.

No – and it does not pretend to. Browser-side face detection is unreliable and often misses cases. The tool gives you an honest manual region selector: you click and drag, and you control exactly which pixels get censored.

Each block averages the pixels underneath it into a single colour. Small blocks (5-10 px) soften details but keep shapes recognisable. Large blocks (60-80 px) destroy nearly all recoverable information within the region.

Yes – that is what Regions mode is for. Draw a rectangle over each area you want hidden; only those rectangles are pixelated. Everything outside stays at the original resolution.

Input: JPG, PNG, WebP, and GIF (first frame only for GIFs). Output: PNG (lossless, transparent-safe) or JPEG (smaller file size, with a quality slider).

No. Each block overwrites the original pixels with a single averaged colour, so the detail under the rectangle is gone for good. Always keep a backup of your source image before censoring.

For faces, use a block size at least one-fifth of the face’s shortest side (e.g. 15 px block for a 75 px-wide face). For text like licence plates or ID numbers, 20-30 px blocks are usually enough to make the characters unreadable at normal viewing distances.

There is no hard cap, but images above roughly 4096×4096 pixels start to feel slow. The pixelation runs in pure JavaScript – very large raw images may take a second or two to render on low-end devices.

Yes. After the page has loaded, every operation happens in-browser. Disconnect from the network and the tool keeps working.

Yes on both counts. Zero uploads – the image never leaves your device – and no sign-up, subscription, or usage caps.