H

h3-studio

Free

MiniMax H3 (Hailuo 3.0) front-end for ComfyUI — local LLM writes H3's structured prompts, live preview, VRAM handoff, native audio.

Visit Website
Related by shared tags and keywordsComfyUIVivijureftl-studio
generative-aiai-videocomfyuicomfyui-custom-nodescomfyui-extensionaiai assistantautomation

Overview

H3 Studio is a ComfyUI custom node that provides a front-end for MiniMax H3 (also called Hailuo 3.0), an open-weights omni-modal model. It supports text-to-video, image-to-video, first-and-last-frame, and reference-to-video generation, producing up to 15 seconds of video at 24 fps with synchronised 32 kHz stereo audio. It installs as a custom node and serves a single page without a second service, CORS setup, or build step.

Because H3 does not accept free text, the tool includes a local LLM prompt assistant that expands a plain brief into H3's structured three-field schema, validates it, and hands it to H3. The assistant can also refine an existing prompt in place. Attached keyframes and subject references are shown to the LLM so that subject definitions describe the actual image.

Additional features include keyframe handling with automatic task selection, subject references that switch to the Ref2VA checkpoint and six-section reference schema, self-correcting prompts, a schema guard, a session filmstrip, format validation and repair, live preview during sampling, progress reporting, automatic VRAM handoff between the LLM and H3, and an endpoint editor.

Requirements include ComfyUI 0.30.0+ with native H3 support, H3 weights from Comfy-Org/MiniMax-H3, and an optional OpenAI-compatible LLM endpoint for the prompt assistant.

Pricing

Free

Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.

Visit Website

Prices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.

Use Cases

  • Teams use H3 Studio to generate video with native audio locally from MiniMax H3.
  • Teams use H3 Studio to create text-to-video, image-to-video, first-and-last-frame, and reference-to-video clips.
  • Teams use H3 Studio to maintain character consistency across clips by attaching subject reference photos.
  • Teams use H3 Studio to expand a plain brief into H3's structured prompt schema using a local LLM.
  • Teams use H3 Studio to drop first and/or last keyframes and have the task follow automatically.
  • Teams use H3 Studio to validate and repair prompt format faults before generation.
  • Teams use H3 Studio to manage VRAM between the LLM and H3 on a single consumer GPU.
  • Teams use H3 Studio to keep session clips and recall them with their prompts and settings.

Features

Generate video from text prompts

Animate images into video clips

Create clips from boundary keyframes

Generate video using subject references

Expand briefs into structured prompts

Validate and repair prompt formats

Preview clips during generation

Recall session clips with their settings

Reviews

No reviews yet. Be the first to share your experience!

FAQ

Why does my MiniMax H3 output have no audio?

Both VAE decodes must feed CreateVideo; the same sampler output goes to VAEDecode with the video VAE and VAEDecodeAudio with the audio VAE. Missing the second produces a silent clip.

What is the correct MiniMax H3 prompt format?

Three labelled fields — integrated_multimodal_description, overall_soundscape, non_diegetic_music — separated by blank lines, in that order. Free text works far less well because ComfyUI has no prompt rewriter.

Does MiniMax H3 support negative prompts?

No. The released checkpoints are guidance-distilled: no negative_prompt, no guidance_scale, one forward pass.