Video Generation
fo-fal renders the catalogue’s video rows marked for fal: MiniMax H3 Max and Seedance 2.5, from a prompt alone or from an image used as the first frame. The finished clip is ordinary project media, added at Motion clips. Video Generation is a separate type from Image Generation even when both share a server or a platform.
Each model’s resolutions and durations come from the catalogue. Reference-to-video is implemented for Seedance 2.5, from reference images. Video-to-video is a taxonomy entry for future work, not a claim about the current renderer.
Subtypes
Section titled “Subtypes”- Text to video — implemented by fo-fal and fo-atlascloud
- Image to video — implemented, from a first frame
- Reference images to video — implemented for Seedance 2.5
- Video to video — future adapters