blurloom is a subscription product. This editor preview has not integrated paid access checks; opening it does not charge you.
The detection engine loads after you select a video
1 Select a video
2 Detect plates and faces
Faces use the YuNet deep learning modelfor improved handling of profiles, angles and lighting. Plates use a fast Haar model, best for clear front-facing plates. Deep models can be slower; watch the progress bar. Missed regions can be added withmanual masksin the next step.
3 Select objects to obscure
Drag to draw a region to obscure
Not masked (click a box to toggle) Selected for masking
4 Process and download
⚠️ Your browser may not fully support MediaRecorder / Canvas capture. For best results, use the latest Chrome / Edge for best results.