A workflow that matches video characters, with video duration equal to or greater than half of audio duration.


Models that need to be downloaded for local deployment (not necessarily using these versions):
1.Wan2_1-I2V-14B-480P_fp8_e4m3fn
https://huggingface.co/Kijai/WanVideo_comfy_fp8_scaled
Place folder: models \ diffusionmodels
2.Wan2_1-InfiniTetalk-Single_fp16.safetensors
Place folder: models \ diffusionmendels \ infinitetalk
https://huggingface.co/Kijai/WanVideo_comfy/tree/main/InfiniteTalk
3.lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16
https://huggingface.co/Kijai/WanVideo_comfy/tree/main/Lightx2v
Place folder: models \ loras