Drop a video and let the AI model identify the subject, remove the background, and replace it with a solid color of your choice. Trim, crop, and preview before exporting.
Ideal for cleaning up talking-head videos, creating content with a clean backdrop, or prepping footage for further compositing. The AI runs entirely in your browser — no uploads, no cloud processing, your video stays private.
The tool runs a deep-learning model (BEN2) directly in your browser using WebGPU or CPU. It analyzes each frame, identifies the foreground subject, and replaces the background with the solid color you pick.
A GPU with WebGPU and FP16 support speeds things up significantly. If your GPU doesn't support it, the tool falls back to CPU processing automatically — it still works, just slower.
The exported video is a WebM file with the background replaced by your chosen solid color. You can use it directly or composite it further in a video editor.
No hard limit, but longer videos need more processing time and memory since every frame goes through AI inference in your browser. Trimming to the section you need first with the [Video Trimmer](/video/trimmer) is recommended for longer clips.