Video to video
Start from footage you already have. Upload a short clip, describe the new look, and the model redraws it — the movement and camera work come from your video, the style comes from your prompt.
How video to video works here
Upload your clip
MP4 or MOV, up to 15 seconds and 32 MB. The model cannot take sound on clips above 24 fps — which is most phone footage, at 30 fps — so we remove the sound for you at upload. A 4K or very high-bitrate file may also go over the size limit; export it at 1080p.
Describe the new look
Say what should change: “a hand-painted watercolor animation”, “a 1990s anime scene”, “the same street at night in heavy rain”. You do not need to describe the motion.
Pay by the second of footage
The price is set by the length of your clip, rounded up to the next second, and shown before you press generate. A 10-second clip is 10 credits.
사양 한눈에 보기
| 모델 | 길이 | 해상도 | 네이티브 오디오 | 참조 이미지 | 시작가 |
|---|---|---|---|---|---|
| Lucy Restyle | 1~15초 | 720p | 아니요 | 사용 안 함 | 5초 기준 5 크레딧 |
Getting a clean restyle
Describe the style or setting, not the action. The motion is already in your clip.
One clear change per render works better than five. Pick the look first, then iterate.
Start with a clip where the subject is large and well lit. Small, fast or heavily compressed subjects are harder to follow.
Trim to the part you need. You pay per second, and shorter clips come back faster.