Open-source AI video creation workspace with an infinite canvas, where a natural-language assistant builds workflows linking text, image and video nodes.
1.7K+7 this week253PythonApache-2.0Updated 2 months ago
AIMixer/ComfyUI_MiniMaxH3_Director
A ComfyUI custom node that adds a multi-segment storyboard director for MiniMax H3 video generation with native stereo audio.
- Stars
- 2,711
- Forks
- 321
- Open issues
- 173
- Last commit
- 25 days ago
What is it?
ComfyUI MiniMax H3 Director is a custom node plugin for ComfyUI that builds on the official MiniMax-H3 nodes. It wraps segment planning, conditioning, sampling, decoding and export into a single director node, so longer videos can be generated as a sequence of timed segments with a visual timeline and per-segment thumbnails.
The plugin supports several task types: text-to-video, image-to-video, first-and-last-frame video, reference-subject video, and video-to-video editing. Audio is generated in the same sampling pass as the picture, producing stereo sound. Optional add-ons include segment-to-segment motion context, a semantic conditioning bridge, a progressive SelfLift first pass, and a refine stage that can re-sample at the same resolution or upscale before a second pass.
The project requires ComfyUI v0.30.0 or later and separately downloadable MiniMax H3 model weights. Installation is via git clone into custom_nodes or ComfyUI Manager, and the repository includes example workflows for each task type.
Key features
- 01Multi-segment timeline with split, even-split, scene-detection and append editing.
- 02Task types cover text, image, first/last frame, reference subject and video-to-video generation.
- 03Native stereo audio generated alongside the video in one sampling pass.
- 04Director pack import and export as zip files containing the timeline and media.
- 05Optional refine stage for same-resolution re-sampling or upscaling.
Use cases
- Generating longer videos as a sequence of timed segments with consistent prompts.
- Creating a video from a first and last frame with transitions in between.
- Keeping a character or subject consistent across several generated shots using reference images.
- Editing an existing source video segment by segment with new prompts.
Quick start
- 1
Upgrade ComfyUI to v0.30.0 or later
- 2
Clone the plugin into ComfyUI's custom_nodes folder
git clone https://github.com/AIMixer/ComfyUI_MiniMaxH3_Director.git - 3
Install the Python requirements
pip install -r ComfyUI_MiniMaxH3_Director/requirements.txt - 4
Restart ComfyUI, then load an example workflow from example_workflows/
Things to know
- You must separately download large MiniMax H3 model weights (UNET, CLIP, and VAEs) before it can generate anything.
- It requires ComfyUI v0.30.0 or later with the official MiniMax H3 nodes.
In collections
Topics
FAQ
- What is ComfyUI_MiniMaxH3_Director?
- A ComfyUI custom node that adds a multi-segment storyboard director for MiniMax H3 video generation with native stereo audio. ComfyUI MiniMax H3 Director is a custom node plugin for ComfyUI that builds on the official MiniMax-H3 nodes. It wraps segment planning, conditioning, sampling, decoding and export into a single director node, so longer videos can be generated as a sequence of timed segments with a visual timeline and per-segment thumbnails.
- Is ComfyUI_MiniMaxH3_Director free to use? What license does it have?
- ComfyUI_MiniMaxH3_Director is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
- Is ComfyUI_MiniMaxH3_Director still maintained?
- Actively maintained. The last commit to ComfyUI_MiniMaxH3_Director was on Oct 13, 2026.
- What can I use ComfyUI_MiniMaxH3_Director for?
- Generating longer videos as a sequence of timed segments with consistent prompts. Creating a video from a first and last frame with transitions in between. Keeping a character or subject consistent across several generated shots using reference images. Editing an existing source video segment by segment with new prompts.
- How do I install ComfyUI_MiniMaxH3_Director?
- Upgrade ComfyUI to v0.30.0 or later Clone the plugin into ComfyUI's custom_nodes folder: git clone https://github.com/AIMixer/ComfyUI_MiniMaxH3_Director.git Install the Python requirements: pip install -r ComfyUI_MiniMaxH3_Director/requirements.txt Restart ComfyUI, then load an example workflow from example_workflows/
- Where can I get ComfyUI_MiniMaxH3_Director?
- The source code and installation instructions are on GitHub: https://github.com/AIMixer/ComfyUI_MiniMaxH3_Director
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A ComfyUI workflow for MiniMax H3 that generates videos from first and last frames, a first frame alone, or text prompts, with a Turbo LoRA and a resolution control panel.
391+202 this week62No licenseUpdated 2 months ago
A local, self-hosted pipeline that turns a novel or script into a long-form video automatically, using ComfyUI, MiniMax H3 and an AI agent.
257+58 this week63JavaScriptMITUpdated yesterday
Self-hosted AI video workspace that turns novels, scripts or product material into characters, storyboards, video clips and CapCut-compatible drafts.
6.2K+253 this week1.2KPythonAGPL-3.0Updated yesterday
Open-source large-scale video generative models for text-to-video, image-to-video, speech-to-video and character animation.
AI & Machine Learning#video-generation#text-to-video#image-to-video#diffusion-model#mixture-of-experts
18.4K+152 this week2.4KPythonApache-2.0Updated 2 months ago
A video generation framework for Diffusion Transformer models that creates videos at flexible resolutions and from images, with training support for baseline and LoRA models.
2.3K+7 this week199PythonApache-2.0Updated yesterday