R2V-04 Character Image + Standalone Audio

RunningHub workflow page: local-only workflow - no RunningHub page in the source material

MiniMax H3 R2V-04 character image + standalone audio

A MiniMax H3 omni-modal video graph exported from episode 288. 34 nodes / 43 links. Four LoadImage slots plus one LoadAudio with TrimAudioDuration; no reference video. 9:16 at 0.4 MP, 10 s.

Saved configuration

• UNETLoader minimax_h3_fl2va_bf16.safetensors; CLIPLoader qwen3vl_32b_minimax_h3_int8_convrot.safetensors (type minimax); VAELoader video minimax_h3_video_vae_fp16.safetensors + audio minimax_h3_audio_vae_fp32.safetensors.
• The saved graphs contain no LoRA nodes.
• Sampling: KSamplerSelect res_multistep with BasicScheduler simple, 20 steps, denoise 1; SamplerCustomAdvanced for the custom chain.
• Conditioning: MiniMaxH3ImageToVideo / MiniMaxH3ReferenceToVideo saved at 1344x768 / 124 / match where present; ResolutionSelector and duration values are quoted per item above; output goes to SaveVideo prefix video/MiniMax_H3.

How to use

This workflow is local-only: the episode's 说明 does not list a RunningHub page for it, so it is distributed as files only.

Limits and not-verified items

No inference was run for this listing: no speed, VRAM, audio-sync or quality claim is made. Values are quoted from the saved nodes. The graph's notes link the official Comfy-Org/MiniMax-H3 model files and the MiniMax blog; those notes are preserved and this listing does not claim to be the original template author. Weights, encoder and VAEs are not bundled; check their licenses separately. The cover image is a promotional still, not a frame produced by this graph.

Tutorial: YouTube · Bilibili. Workflows supplied through AI-KSK; verify workflow, node, model, image/audio and cover rights separately. No blanket commercial or redistribution license is asserted.