Drag the image to choose the crop. The frame automatically matches the selected aspect ratio.
Add any mix of references. Tags follow ComfyUI's presentation order: all images first; for each video, its embedded soundtrack is tagged as <Audio n> before the video as <Video n>; standalone audio tags come last. A video's soundtrack is paired automatically when present.
No references uploaded.
No references uploaded.
No references uploaded.
FL2VA and Ref2VA keep their existing default checkpoints. Optional DaSiWa Hybrid V2 can be selected for either mode; its matching Turbo LoRA supports 4–8 steps (8 recommended).
Your video will appear here
Generation can take several minutes on supported hardware.