TencentHunyuanVideo-Foley
HunyuanVideo-Foley
HunyuanVideo-Foley generates audio conditioned on video. Its purpose is to align sound with visible events and the scene’s meaning.
Model design and access
The model combines multimodal and unimodal transformer processing. It is an audio generation stage for existing video, not a replacement for a video generator. The repository supplies model weights and an interactive inference workflow.HunyuanVideo-Foley official model documentation
Practical use
Practical guidance. Listen with and without the picture. Check event timing, background continuity and whether the model invents a sound that changes the apparent meaning of the scene. Keep generated effects on a separate track for editing.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| License | Tencent Hunyuan Community License; custom terms applyHunyuanVideo-Foley official model documentation |