<h1 id="anikin-labs-minimax-h3-ref2va-reference-generation-v2">Anikin Labs — MiniMax H3 Ref2VA Reference Generation v2</h1>Create entirely new video scenes from a reference image using MiniMax H3 in ComfyUI.Unlike traditional Image-to-Video workflows, Ref2VA uses your input image as a visual reference rather than requiring it to become the opening frame.Upload a character reference, describe a new location, action, camera movement, and audio — and let MiniMax H3 generate a new video while attempting to preserve the character's visual identity.<h2 id="features">Features</h2><ul><li>Reference-to-Video-and-Audio (Ref2VA) generation</li><li>Create new scenes from a character reference image</li><li>Turbo mode: 8 sampling steps using a Ref2VA-specific acceleration LoRA</li><li>Normal mode: 20 sampling steps without the Turbo LoRA</li><li>Convenient Normal / Turbo mode switch</li><li>3 additional optional LoRA slots, bypassed by default</li><li>Adjustable resolution and aspect ratio</li><li>Adjustable video duration</li><li>Video and audio generation in a single workflow</li><li>Support for additional reference inputs through the MiniMax H3 Ref2VA node</li><li>Integrated download links and English setup instructions</li></ul><h2 id="how-to-use">How to Use</h2><ol><li>Download the required models using the links included inside the workflow.</li><li>Place each model in its corresponding ComfyUI folder.</li><li>Load your reference image in the Character Reference — Picture 1 node.</li><li>Edit the prompt to describe the new scene, character action, camera movement, and audio.</li><li>Select your desired aspect ratio, resolution, and duration.</li><li>Choose Turbo or Normal mode.</li><li>Run the workflow and review the generated video.</li></ol><h2 id="generation-modes">Generation Modes</h2>Turbo — 8 StepsUses the LightX2V Ref2VA Turbo 8-step LoRA for faster generation.Normal — 20 StepsRuns the base Ref2VA model without the Turbo acceleration LoRA. Useful for quality comparisons and more demanding scenes.The workflow includes three additional optional LoRA slots for experimentation. Enable only compatible LoRAs and test them individually.<h2 id="requirements">Requirements</h2>Suggested minimum (experimental):<ul><li>NVIDIA GPU with 12 GB VRAM</li><li>32 GB system RAM</li><li>SSD storage</li><li>Compatible ComfyUI installation</li></ul>Recommended:<ul><li>NVIDIA RTX GPU with 24 GB VRAM</li><li>64 GB RAM or more</li><li>NVMe SSD</li></ul>Lower-memory systems may require model offloading. Performance and memory requirements depend on resolution, duration, and installed components. The suggested minimum is not a guaranteed tested configuration.<h2 id="important-notes">Important Notes</h2><ul><li>This is not a character-swap workflow for an existing video.</li><li>Reference images guide generation but do not guarantee perfect identity consistency.</li><li>Complex backgrounds, movement, fine details, and facial consistency may vary between generations.</li><li>Audio and dialogue quality depend on the scene and prompt.</li><li>Required model weights are not bundled with the workflow JSON.</li><li>Model downloads and installation paths are listed inside the workflow.</li></ul><h2 id="credits">Credits</h2>Workflow configuration and documentation: Anikin LabsBase model: MiniMax H3 ComfyUI model distribution: Comfy-Org Turbo acceleration LoRA: LightX2VThis workflow builds on the original model, official ComfyUI nodes, and acceleration LoRA created by their respective developers.<hr />Anikin Labs — Exploring Local AI, ComfyUI, and Creative Workflows.

MiniMax H3 Ref2VA Reference Generation v2 (8-Step Turbo / 20-Step / 3 LoRA Slots) (Anikin Labs)

Initial Public Release — v1.0<ul><li>MiniMax H3 Ref2VA video and audio generation.</li><li>Turbo 8-step and Normal 20-step modes.</li><li>Three optional LoRA slots.</li><li>Adjustable resolution, aspect ratio, and duration.</li><li>English documentation and direct download links for all required models.</li><li>Clean reference input and reusable prompt template.</li></ul>

Files with SHA256: d4a4f3ea84f58233466738f02c65a37c9f7dcd61643bb89e44cc1eff2acb0a52