MiniMax H3 Chain Director V1.0 chains multi-segment video generation in ComfyUI
Far_Patience4517 · reddit · 2026-10-24
The author's MiniMax H3 Dual Segment workflow has evolved into Chain Director V1.0, a custom Director node controlling the full generation pipeline from one panel.
Key features:
- Single and Chain modes: standalone segments or longer multi-segment sequences;
- General + per-segment prompts with shared project-wide instructions;
- Video continuity: MiniMax H3 Motion Context carries audiovisual context between segments for consistent motion, scenes, and transitions;
- Automatic context trimming of repeated transition frames;
- Independent BASE and UPSCALE stages with separate models/LoRAs/samplers/steps, plus configurable 3D latent upscaling with a second sampling pass;
- One-panel control of duration, resolution, seed, sampler, and more; live previews via KJNodes; external image/video/audio references; final export with audio.
The public workflow is sanitized; download and full docs are on Civitai. The author is collecting feedback across different ComfyUI setups.
More from Multimodal
- cocktailpeanut: without realtime generation, a world model is just a video model with WASD prompts — cocktailpeanut · 2026-10-24
- Flowing water and lava come to 3D Gaussian Splatting scenes — jonstephens85 · 2026-10-24
- Keyframe Labs doubles avatar model resolution, claims 180ms latency at $0.06/min — ycombinator · 2026-10-24
- ElorianAI's Glimpse multimodal model IDs Ed Sheeran concert from blurry video, claims to beat ChatGPT at fraction of cost — AndrewDai · 2026-10-24
- Letting Claude run 10+ hours to auto-build full ComfyUI workflows — ThatKnarfGuy · 2026-10-24
- Ideogram Vector + Opus 5.5: one prompt to editable SVG, then a full animated explainer — CSProfKGD · 2026-10-24