Why pixelate.video exists

Before a video goes online, someone often has to hide something in it: a bystander's face, a child at a school event, a car's number plate, a colleague's name on a screen. Most online blur tools ask you to upload the very footage you are trying to protect. pixelate.video was built so that you never have to: the censoring happens on your own device.

Built for privacy work

The editor is used by people who share dashcam and doorbell clips, parents posting family videos, journalists and researchers anonymising interviews, and teams redacting screen recordings before a demo. It offers three effects, pixelation, blur and a solid black box, and lets you apply them to the whole frame or to as many boxes and ovals as you need, each with its own start and end time.

How detection and tracking work

Face finding uses Google's MediaPipe face detector (the BlazeFace short-range model), which is served from this site and runs inside your browser. Every face it finds becomes an oval that follows the person. For other objects, the tracker compares a small template of the area against each frame and follows the best match, adapting as the object changes size or lighting. Decoding and encoding use the WebCodecs API and the open-source mediabunny library. None of these steps needs a server.

Limits worth knowing

• Tiny, distant or turned-away faces can be missed, so always watch the export before you share it.
• Light blur and small pixel blocks can leave text and faces readable. For passwords, IDs and plates, use the black box or large blocks.
• Very long or high-resolution files are limited by your device's memory and speed.

Our principles

No uploads, no account, no watermark and no charge. pixelate.video uses no cookies and counts visits only in aggregate; the privacy page lists exactly what is measured. Start with the guides on blurring faces, licence plates or censoring a video.