Choose a clear source video
Use footage where the subject is large enough to see, remains in frame, and is not hidden by foreground objects.

The reliable workflow is simple: start with a clean source, choose whether you need transparency or a replacement image, inspect difficult frames, and export in the format that preserves the result.
Use footage where the subject is large enough to see, remains in frame, and is not hidden by foreground objects.

Choose transparent WebM when you will composite later, or upload an image when the replacement should be baked into the result.

Check hair, hands, motion blur, shadows, and moments where the subject overlaps similar colors.
Preview transparent output over both light and dark backgrounds, or confirm that a replacement scene matches the intended crop.
Even lighting, visible edges, and a steady camera help the foreground model. You do not need a green screen, but heavy compression and rapid motion make the mask harder to stabilize.
Transparent WebM gives you the freedom to reuse the foreground in different scenes. A precomposited background is easier to publish where transparency is not supported.
Start from the required final asset: a transparent overlay, a clip composited over a still image, or foreground footage for another editor. That choice determines the output format and the checks you need to perform.
The fastest workflow is not the one with the fewest clicks; it is the one that produces a file your next tool can actually use without another round of masking, conversion, or edge repair.
Use the highest-quality original available instead of a copy downloaded from a social platform. Keep the subject large enough to see, avoid aggressive compression, and prefer even lighting with visual separation between foreground and background.
Preview the beginning, middle, and end before upload. A clip that looks easy in its first frame may contain crossed hands, hair movement, camera cuts, fast rotation, or an object that briefly disappears behind another element.
The live remover accepts MP4, MOV, and WebM. Transparent mode returns WebM; custom-background mode accepts one JPG, PNG, or WebP image up to 10 MB. A still image is composited behind the isolated foreground—the product does not currently accept a replacement video.
The free trial supports one verified clip up to 5 seconds and 50 MB. Paid plans raise the limit to 60, 180, or 300 seconds and 200 or 500 MB depending on tier. Processing credits are rounded up in five-second blocks, so a 6-second clip uses 10 credits.
Yes. The online workflow handles upload, AI segmentation, and export in the browser.
No. A clean, well-separated subject helps, but the model does not rely on a selected green color.
Use a higher-quality source, reduce motion blur where possible, and check whether the subject becomes too small or obscured in those frames.
Credits follow processed duration and round up to the next five-second block. A clip from 1–5 seconds uses 5 credits, 6–10 seconds uses 10 credits, and the same rule continues for longer clips.
The verified free trial allows one clip up to 5 seconds and 50 MB. Depending on the paid tier, limits increase to 60, 180, or 300 seconds and 200 or 500 MB. Background images are limited to 10 MB.