Wan 2.1 LoRA video-to-video redraws a clip you already have from a prompt, keeping its motion and timing, with up to three of your own LoRA files on Alibaba's 14B model. Output at 480p or 720p in the shape of the source, nothing filtered.
Workbench
These examples were written for this endpoint. Replace the prompt and sample media URLs with your own accessible inputs, then check the current quote and model availability before submitting.
- Output5s
gondola-candlelit-grotto No LoRA, 480p, `strength` `0.95`, on the five-second gondola clip. At this strength the source keeps only its composition and camera move — a woman reclining in a boat, a slow push-in — and the prompt supplies everything else, so the brief asks for exactly that: a candlelit sea cave, a gilded boat and a gold silk gown. Lower strengths hold the original people and set and change very little.
- 211014
- 0.95
- 480p
- Output5s
gondola-cloak-restyle The same clip and prompt at 480p (800 x 448). Again a photographic redraw rather than a watercolour, this time with a velvet cloak and loose curls. Neither run set a seed, and each one redrew her face and wardrobe differently.
- 0.9
- 480p
- 5s
- Output5s
mooring-mask-portrait A flat-colour LoRA at 1 with its trigger words on a portrait clip (448 x 832 at 480p from a 688 x 1248 source). The man, the mooring pole and the raised mask carry over and the scene is redrawn in cooler, sharper light, but the clip stays live action: the LoRA that turns text-to-video into cel animation did not do so on footage at strength 0.9.
- 211001
- 0.9
- 480p
This example starts at $0.24 per request. The Playground updates the estimate when you change the settings.