MiniMax H3 RefMod in ComfyUI: Build a Reusable Character, Motion, Voice and Scene Library

Description and use

Six ComfyUI workflows enable creation and reuse of pre‑encoded RefMod assets for character, motion, voice, and scene, saving time and resources.

Problem

Repetitive feeding of original media into every generation graph wastes time and storage.

Outcome

Six ComfyUI JSON workflows enable creation of reusable RefMod assets and their application to generate new video clips without re‑encoding source media.

Workflow

• Create character RefMod from image folder
• Create motion RefMod from video file
• Create audio RefMod from audio clip
• Create scene RefMod from two images
• Generate video A by applying four saved refs
• Generate video B by text‑encoding three refs plus a character

Evidence

These workflows extract reusable, pre‑encoded reference assets and load them when rendering a new clip.

Source

The generator graphs select the following exact checkpoint filenames, not an interchangeable generic “H3 model”

Source

If nodes appear missing, resolve the node packages and restart/reload ComfyUI; do not substitute similarly named nodes from unrelated projects without checking their inputs.

Source