How to use Blurwise, step by step

Add your files

Drag your videos and photos into the list, or click Add files — you can mix clips and stills freely, and HDR videos are fine: their colours are converted automatically. Choose an output folder and a Video quality (Standard suits most uses; High keeps more detail in larger files), and leave “Write an anonymisation record” on if you may need to prove later what was anonymised. Then press Find faces.

• Sensitivity sets how many faces Blurwise looks for: Thorough also catches turned or moving faces, Balanced suits most videos, Precise only takes clear ones. The exact value is under Advanced.
• Something not behaving as expected? Restore default settings puts the remembered detection and review settings back to their defaults.
• Pro and Studio process several videos in one pass; Free works on one video at a time, with as many photos as you like.

Let it find the faces

Blurwise scans each file with an on-device neural network, finding and tracking every face frame by frame. It all runs on your computer — nothing is uploaded. You can pause or stop anytime, and results are cached so going back never re-detects.

Choose how to cover each face

Per face, pick a cover style: a rectangle, an oval, a thin strip over just the eyes, or an emoji / your own image laid on top. Whatever you choose, the face is always blurred underneath — so nothing shows through if an image has transparent areas or the box shifts.

• Rectangle and oval are available on every version.
• Eye strip, emoji and custom images need Pro or Studio.

Change many faces at once

In a crowd, pick the faces to change together: Ctrl+click rows one by one, or Shift+click to take every row between two. Switch blur on or off, or choose a cover, on any of them and the change applies to all of them. A plain click or Esc clears the selection.

• “Show” above the list narrows it to the faces to check, the blurred ones or the ones left visible.
• Faces marked “Check” deserve a second look: the detector was unsure, or saw them only briefly.

Follow an object

Number plates, badges, documents, screens — anything the face detector won’t catch. Press Track, draw a box around it, adjust it until it fits and press Start: Blurwise follows it frame by frame, and the box grows and shrinks with the object. If it loses the object, the box holds its last position so nothing slips into the clear until you say it leaves the shot.

Disguise voices and silence audio

A blurred face still leaks identity through the voice. Under the video, “Add voice disguise” marks a stretch at the playhead and disguises it irreversibly, and “Add silence” mutes one — for a name, address or phone number said out loud. Every stretch is listed in the Audio tab: move its edges with Start here and End here, remove it with ✕, and give each speaker their own tone, from Much lower to Much higher, so several people can still be told apart. “Whole track” disguises the entire recording.

• Normal playback plays the original voice: press ▶ on a stretch, or Listen, to hear it exactly as it will export.
• Turn on “Play the sound while the video plays” to follow the footage by ear; silenced stretches stay silent.
• Every tone is equally irreversible — the disguise is just as strong whichever you pick.

Change your mind after exporting

Finished files turn green in the list. Spotted one more face while watching the result? Select the file and press Review again (or right-click it): it goes back into the list, and your per-face choices, hand-drawn boxes and audio stretches come back exactly as you left them, with nothing detected again. The new export is saved beside the first one, never over it.

• If you change a detection setting first, the faces are found again: your blur look, margins, audio stretches and hand-drawn boxes carry over, and each face starts blurred.