Remove the background from a video using the RVM (Robust Video Matting) AI model (github.com/PeterL1n/RobustVideoMatting). The model runs right in your browser via ONNX Runtime Web (WebGPU, with a WASM fallback), frame-by-frame, building an alpha mask that is merged back over the clip so only the subject remains. RVM is trained for human matting, so it works best on videos with people — other subjects may not cut out cleanly. Everything runs locally in your browser using WebCodecs — your file is never uploaded to a server, so it stays private and conversion starts instantly.
How to remove background
- Drop a video onto the box above.
- Click Remove background — the RVM model processes each frame on your device.
- Preview the cutout, then click Create MP4 to download the result.
Frequently asked questions
- How does the background removal work?
- The RVM (Robust Video Matting) neural network — github.com/PeterL1n/RobustVideoMatting — segments the foreground from the background on each frame, producing an alpha mask that is merged back over the clip.
- What works best?
- RVM is trained for human matting, so it works best on videos with a person in them. Other subjects (objects, animals, scenery) may not separate cleanly.
- Is my video uploaded?
- No. The RVM model runs in your browser via ONNX Runtime Web (WebGPU, or a WASM fallback), with WebCodecs + WebGL for decoding and compositing — the file never leaves your device.
- Why is processing slower than other tools?
- Background removal runs a neural network on every frame, which is heavier than a normal re-encode. Speed depends on your device and the clip length.