Back to models
Alibaba
Video modelWan 2.7
Wan 2.7 is a video generation model in Alibaba Cloud's Wan family, offering text-to-video, image-to-video, and reference-based creation. It suits advertising concepts, product demonstrations, character clips, and stories spanning multiple shots. Start with text to plan characters, settings, and camera work. Existing images can define an opening frame or both the opening and ending frames, while video continuation extends an existing clip. Reference generation combines images, video, and voice references to carry character appearance, scene settings, and voice characteristics into new content. Audio can be generated to match the visuals, and supported modes can use supplied audio to guide performance and timing. Creators can explore a visual direction first, then add material to refine characters, movement, and sound.
- Multimodal References
- Synchronized Audio and Video
Key features
Multimodal character references
Combine images, video, and voice references to carry characters and settings into new clips.
Frame control and video continuation
Guide generation with an opening frame or opening and ending frames, or continue an existing video.
Coordinated sound and shots
Supports automatic audio and multi-shot storytelling, with supplied audio guiding performance timing in supported modes.
