First and last frame control
Give the model where a shot starts and where it ends, and it interpolates the motion between them instead of guessing at an ending.
Alibaba
Next-generation AI video by Alibaba. Native 1080p at up to 15 seconds, with instruction-based editing and character references that hold a face, an outfit and a voice across every shot.
Features
Give the model where a shot starts and where it ends, and it interpolates the motion between them instead of guessing at an ending.
Restyle, replace an element or swap a background by describing the change. The rest of the frame stays where it was.
Multi-subject and multi-shot references keep the same person — face, wardrobe and voice — across an episode rather than a single clip.
Feed up to nine reference images in one pass to fix wardrobe, props and set dressing before the first frame renders.
Examples
A woman walks along the shore at dawn, soft golden light, cinematic
Try this promptReplace the pomegranates with apples
Try this promptChange background to snowy mountain at dusk. Keep subject identical.
Try this promptRestyle the outfit every second: streetwear, red dress, yellow trench. Same pose, no warping.
Try this promptThese prompts are reproduced from Alibaba's own Wan 2.7 material. Run them in the Studio to see your own render.
Workflows
Cost is charged per second of output, so a 15-second take costs three times a 5-second one at the same resolution. Credits come from one shared balance across image, video and speech, and the exact estimate is shown in the Studio before you generate.
Built for
FAQ
Limits, audio, licensing and price — the things worth knowing before you spend a credit.
Open the Studio, pick Wan 2.7, and see the credit estimate before you spend anything.