MiniMax H3 Singularity Reference To Video ComfyUI Workflow
Create coherent videos from character, style, and scene references with stable subjects, controlled motion, and sharper detail.
AI model workflow
MiniMax H3 ComfyUI Workflows
Create video from text, still images, or visual references with MiniMax H3 in ComfyUI. Explore synchronized audio, first-and-last-frame transitions, and dedicated workflows for talking or singing characters. Run them on RunComfy Cloud, download workflow JSON, or deploy supported workflows as APIs.
20 workflows
Featured workflow
Minimax H3 I2V Dual Clock 8-Step| Ultra-Fast High-Quality Video
The workflows
20 MiniMax H3 ComfyUI Workflows
Compare MiniMax H3 ComfyUI examples by prompt, image, or reference input, then check audio and frame-control support. Each detail page explains its specific setup and acceleration options. Run it on RunComfy Cloud, download the workflow JSON, or deploy supported workflows as APIs.
VDN H3 ComfyUI Text To Video | Faster MiniMax H3 VDN 8-Step Workflow wihout LoRA
MiniMax H3 VDN makes cinematic video with audio in just 8 steps.
MiniMax H3 Latent Upscaler | BUNNYH3 Two-Pass Workflow
Generate higher-resolution videos from text or reference images with built-in upscaling and second-pass refinement.
MiniMax FastH3 4-Step
MiniMax FastH3 four-step video drafts: Fast cinematic clips from text or images.
MiniMax H3 Fun Control Image-to-Video
Create controlled reference-image-to-video clips with MiniMax H3 Fun ControlNet, using depth, pose, and optional SolAttn branches.
MiniMax H3 + Music 3 Music Video
MiniMax H3 and MiniMax Music 3 pipeline turns any song into synced cinematic scenes.
MiniMax H3 Dual-Sampling Reference-to-Video
MiniMax H3 makes consistent cinematic videos with synced audio - fast.
MiniMax H3 3D Rendering Video
MINIMAX H3 turns 3D clay renders into cinematic videos.
MiniMax H3 Character Replacement
MiniMax H3 swaps characters while keeping every move and shot intact.
MiniMax H3 Dual Clock 8-Step I2V
Minimax H3 Dual Clock 8-Step turns one image into synced video and stereo sound fast.
MiniMax H3 First-Last Frame Turbo
Turbo H3 turns first and last frames into videos with stereo audio.
MiniMax H3 Dual-Character Talking & Singing (4-Step)
Two reference characters talk and sing face-to-face with lip sync in 4-step acceleration.
MiniMax H3 Multi-Reference Video (4-Step Turbo)
MiniMax H3 4 Steps Turbo turns reference photos into consistent videos with stereo sound.
MiniMax H3 Image-to-Video (4-Step Turbo)
Hailuo H3 Turbo turns one photo into synced video and stereo audio.
MiniMax H3 Text-to-Video (4-Step Turbo)
Hailuo H3 Turbo makes videos with stereo sound in four steps.
MiniMax H3 Talking Avatars
Turn portraits into powerful talking avatars with cloned voices and synced lips.
MiniMax H3 All-in-One Multi-Reference Video
FAST MiniMax H3 turns nine references into cinematic video with stereo sound.
MiniMax H3 Reference-to-Video (Stereo)
MiniMax H3 Ref2VA creates cinematic video with native stereo audio.
MiniMax H3 Image-to-Video
MiniMax H3 turns one image into cinematic video with synced audio.
MiniMax H3 Text-to-Video
MiniMax H3 makes cinematic videos with synced stereo audio from one prompt.
20 of 20 workflows shown
About MiniMax H3
What can MiniMax H3 do in ComfyUI?
MiniMax H3 in ComfyUI supports text-to-video, image-to-video, and reference-to-video generation. Audio-enabled workflows add synchronized sound, while reference images and opening and ending frames help guide the subject and composition of a shot.
Choose text-to-video to build a scene from words, image-to-video to animate a still, or reference-to-video when the subject and visual direction should follow supplied images. Dedicated performance setups handle talking or singing characters.
Acceleration, character replacement, and music-video assembly are choices within specific workflows, not features included in every MiniMax H3 setup. For example, the music-video pipeline combines MiniMax H3 with MiniMax Music 3. Open a matching example for its inputs, supporting models, and settings.
References and model resources: MiniMax H3 GitHub, Explore MiniMax H3 models on Hugging Face, Explore MiniMax H3 models on Civitai.
Capabilities
Explore MiniMax H3 capabilities in ComfyUI
Choose the source of your video idea: text, a still image, visual references, or opening and ending frames. Dedicated setups also support character performances.
Text-to-video with audio
Create a scene from a written prompt using a video-and-audio generation setup.
Image-to-video with audio
Use a still image as the starting point for motion and synchronized sound.
First-and-last-frame video
Provide an opening and ending image to guide the transition between them.
Talking and singing characters
Create character performances with a setup designed for speech or singing.
MiniMax H3 on RunComfy
Create with MiniMax H3 in a cloud ComfyUI workspace
Open a prepared MiniMax H3 example and adjust the inputs and video controls available in its ComfyUI setup. Use the cloud environment for generation, or download the JSON and follow the detail page's requirements for local use.
Run a MiniMax H3 workflow nowEditable scene inputs
Work with the prompts, images, and reference controls included in the selected example.
Reference and audio assets
Supply the reference images or audio assets required by the video result you want.
Prepared dependencies
Start with the dependencies used by the example, including additional models where its pipeline needs them.
Cloud GPU options
Choose hardware for the video size and processing stages in your selected setup.
Applications
What you can make with MiniMax H3
Find a production example for music videos, animated renders, character scenes, or talking avatars. Combined pipelines list the additional models they use.
AI Music Videos
Build cinematic song visuals with a pipeline combining MiniMax H3 and MiniMax Music 3.
Character Replacement
Replace a character in footage with a dedicated reference and masking pipeline.
Talking Avatars
Develop a speaking portrait for an explainer, social clip, or character scene.
Frequently asked questions
MiniMax H3 in ComfyUI: FAQ
Understand video input options, check audio and character support, and choose a suitable cloud, local, or LoRA training path.
What is MiniMax H3 used for in ComfyUI?
MiniMax H3 is used to create video from text, still images, or visual references in ComfyUI. The examples include audio-enabled scenes, controlled transitions, and character performances. Choose by the input you already have and the type of shot you need.
Should I use MiniMax H3 text-to-video, image-to-video, or reference-to-video?
Use text-to-video when you want to describe a new scene, image-to-video to animate a still, and reference-to-video to guide a subject or visual direction. First-and-last-frame control is useful when the clip needs to begin and end with specific images.
Does MiniMax H3 generate native stereo audio?
Some examples include native stereo audio, including the MiniMax H3 text-to-video setup. Check the selected example's outputs and audio inputs. Generated sound, a supplied speech track, and a separate music model are different audio paths.
Can I control the first and last frame with MiniMax H3?
Yes. The first-and-last-frame example accepts opening and ending images to guide a transition. Choose this when both endpoints matter, rather than relying on a single starting image to determine the ending.
Can MiniMax H3 make talking avatars or singing characters?
There are dedicated examples for a talking avatar and talking or singing characters. Use their required references and performance inputs. A general video-generation setup may not expose the same controls.
Can I train or use a MiniMax H3 LoRA?
Yes. RunComfy Trainer supports MiniMax H3. Use the LoRA with a compatible MiniMax H3 workflow and keep the same model family between training and inference.
Where can I download a MiniMax H3 ComfyUI workflow JSON?
Use Download JSON on a MiniMax H3 card or detail page and open it in ComfyUI. Follow that detail page for the required models, custom nodes, and input assets. Downloading the JSON alone does not install the local environment.
Related resources
More ways to use MiniMax H3
Train a MiniMax H3 LoRA, or use RunComfy Models pages for MiniMax H3 and Hailuo when you want direct video model generation.
MiniMax H3 Models
Use MiniMax H3 model generation directly when you do not need a full ComfyUI workflow graph.
Open MiniMax H3 ModelsMinimax Hailuo Models
Explore adjacent Hailuo video models when your goal is direct video generation rather than workflow editing.
Open Hailuo ModelsTrain a MiniMax H3 LoRA
Create a custom MiniMax H3 LoRA for a character, style, product, or video look.
Open MiniMax H3 Trainer