Make a second H3 clip from the last frame of the first, then join both, picture and sound.
Tested workflow plus a sample first frame.
Watch the video: https://www.youtube.com/watch?v=i3Zv6QxFgPg
What you need
- ComfyUI 0.30 or newer
- minimax_h3_fl2va_pruned_int8_convrot.safetensors (19.5 GB) in models/diffusion_models
- qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors (14.6 GB) in models/text_encoders
- minimax_h3_video_vae_fp16.safetensors (4.9 GB) in models/vae
- minimax_h3_audio_vae_fp32.safetensors in models/vae (without it the clip has no sound)
- A graphics card with 24 GB or more (under that it is slow or may not run)
How to use it
- Drag the workflow into ComfyUI.
- Upload the included first frame or your own picture. For your own, match its shape to the clip size.
- Pick the model files in the loaders, keep the CLIP loader type on minimax.
- Edit the two prompts and Run. The output joins both clips and their sound.
- Watch the join. If the look jumps, bring the two prompts closer and run again.
The step-by-step guide, the video and more free workflows are on my portal.

