Tencent-Hunyuan/HunyuanCustom
A multimodal diffusion-model video generation system for customized video creation from audio, images, and other video inputs.

Velocity · 7d
+3.1
★ / day
Trend
→steady
star history
HunyuanCustom is Tencent’s video generation architecture built on diffusion models that enables customized video creation through multiple input modalities. It supports audio-driven video generation, image-to-video conversion, and video-driven editing. The project releases inference code and pre-trained model weights, and has been integrated into ComfyUI for wider accessibility.