Four new ways to get seen and get paid: sticker placement, remix galleries, cosmetic packs and collab collections. Plus free placements on every image, so you can join in without spending a thing.

The Creator Collab Update

You can gift a membership to someone else, or buy one with Buzz instead of a card. Here's exactly what each one gets you.

Membership Update

Sell a set of cosmetics as one listing at one price, instead of one piece at a time. Two to twenty items, any kinds mixed, other creators' work included.

Cosmetic Packs

We're a Day-0 launch partner for MiniMax H3, a next-gen open-weights multimodal video model - and it's already in the generator. No waitlist. Open weights land in a few days.

MiniMax H3 is live: open-weights video, generate now

MiniMax H3 Original MP3 Lip-Sync Workflow Fixes the issue where input audio is regenerated into pseudo-Chinese vocals. This workflow passes through your original MP3/vocal directly to the final video output while maintaining H3's precise lip-sync.<ul><li>Generation Speed: ~228s for 10-second video (INT8/NVFP4 environment)</li><li>No missing custom nodes needed, standard setup.</li><li><div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist src="https://www.youtube.com/embed/RIN_mp17EBQ" start="0"></iframe></div></li></ul>

# MiniMax H3 Turbo Workflow (Reference to Video + Audio Lip-Sync + Upscalers)This is an optimized ComfyUI workflow for MiniMax H3 Reference-to-Video enhanced with Turbo LoRA and integrated Upscaler pipelines (4x-AnimeSharp &amp; RTX Video Super Resolution).Generated 15s high-quality character dancing/singing videos in ~180s - 200s (RTX 4090 / 5090) with consistent character appearance!---## 📺 YouTube Tutorial &amp; Comparison ReviewWatch the full setup tutorial, comparison with LTX 2.3, and upscaler performance benchmarks:👉 [Insert Your YouTube Video Link Here]---## ✨ Key Features &amp; Improvements1. Turbo LoRA Acceleration: Reduces sampling steps from 20+ down to 6-8 steps <code>res_multistep</code>), cutting generation time by more than 50%.2. Audio Lip-Sync &amp; Music Video Support: Fixes previous audio replacement issues. Simply drag-and-drop your custom MP3/WAV into the workflow.3. Built-in Upscaling Pipelines: - RTX Video Super Resolution (VSR): Ultra-fast real-time AI upscaling, ideal for rapid iterations and previews. - 4x-AnimeSharp + ImageScale: Supreme line-art reconstruction and edge clarity for final masterpieces.4. Strong Character Consistency: Reference image identity is preserved throughout the 15-second performance.---## 🛠️ Required Models &amp; Custom Nodes### 1. Base Models (MiniMax H3)Place in their respective standard ComfyUI folders <code>models/unet</code>, <code>models/clip</code>, <code>models/vae</code>):- UNet: <code>minimax_h3_ref2va_pruned_int8_convrot.safetensors</code> (or fp8/fp16)- Text Encoder: <code>qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors</code>- Video VAE: <code>minimax_h3_video_vae_fp16.safetensors</code>- Audio VAE: <code>minimax_h3_audio_vae_fp32.safetensors</code>### 2. Turbo LoRA- <code>minimax_h3_turbo_v4_step600_ema_pruned_comfyui.safetensors</code> - Download from: [Hugging Face (drbaph)](<a target="_blank" rel="ugc" href="https://huggingface.co/drbaph/MiniMax-H3-Turbo-Lora-ComfyUI/tree/main">https://huggingface.co/drbaph/MiniMax-H3-Turbo-Lora-ComfyUI/tree/main</a>) - Place into: <code>ComfyUI/models/loras/</code>### 3. Custom NodesRun the following in <code>ComfyUI/custom_nodes/</code>:```bashgit clone [<a target="_blank" rel="ugc" href="https://github.com/Larryvrh/ComfyUI-MiniMax-H3-Turbo](https://github.com/Larryvrh/ComfyUI-MiniMax-H3-Turbo)">https://github.com/Larryvrh/ComfyUI-MiniMax-H3-Turbo](https://github.com/Larryvrh/ComfyUI-MiniMax-H3-Turbo)</a>git clone [<a target="_blank" rel="ugc" href="https://github.com/Larryvrh/ComfyUI-H3-MultiShot](https://github.com/Larryvrh/ComfyUI-H3-MultiShot)">https://github.com/Larryvrh/ComfyUI-H3-MultiShot](https://github.com/Larryvrh/ComfyUI-H3-MultiShot)</a>

v1.3

This workflow is a ultra-fast MiniMax H3 Reference-to-Video pipeline accelerated by Turbo LoRA (6~8 Steps). It enables you to generate a 15-second high-quality reference video with lip-sync in under 200 seconds (tested on RTX 5090). (MiniMax H3のReference-to-VideoをTurbo LoRAで高速化し、15秒動画を約200秒弱で生成できる爆速ワークフローです。)【Required Models / 必要なモデル】<ul><li>Reference Diffusion Model:<ul><li><code>minimax_h3_ref2va_pruned_int8_convrot.safetensors</code> (or fp8 version)</li><li>Save to: <code>ComfyUI/models/diffusion_models/</code> (or <code>models/unet/</code>)</li></ul></li><li>Turbo LoRA Model:<ul><li><code>minimax_h3_turbo_v4_step600_ema_prun...safetensors</code></li><li>Save to: <code>ComfyUI/models/loras/</code></li></ul></li><li>Text Encoders &amp; VAE:<ul><li>Standard MiniMax H3 CLIP/Text Encoders in <code>models/clip/</code> and VAEs in <code>models/vae/</code>.</li></ul></li></ul>【Custom Nodes Installation / カスタムノード導入手順】Open your command prompt in the <code>ComfyUI/custom_nodes</code> directory and run the following commands: (ComfyUIの <code>custom_nodes</code> フォルダを開き、アドレスバーに <code>cmd</code> と入力してEnterを押した後、以下のコマンドを実行してください)command1：git clone <a target="_blank" rel="ugc" href="https://github.com/Larryvrh/ComfyUI-MiniMax-H3-Turbo">https://github.com/Larryvrh/ComfyUI-MiniMax-H3-Turbo</a>command2：git clone <a target="_blank" rel="ugc" href="https://github.com/jlucasmcrell/ComfyUI-H3-Multishot">https://github.com/jlucasmcrell/ComfyUI-H3-Multishot</a>Restart ComfyUI after cloning. (クローン完了後、ComfyUIを再起動してください)【Recommended Settings / 推奨設定】<ul><li>Sampler: <code>MiniMax-H3 Turbo Sampler</code></li><li>Scheduler: <code>simple</code></li><li>Steps: <code>6</code> - <code>8</code></li><li>LoRA Strength: <code>1.0</code></li><li>Duration: <code>15.0</code> seconds (at 24 FPS)</li></ul>【Credits】<ul><li>MiniMax H3 Turbo Node: <a target="_blank" rel="ugc" href="https://github.com/Larryvrh/ComfyUI-MiniMax-H3-Turbo">Larryvrh</a></li><li>MultiShot Node: <a target="_blank" rel="ugc" href="https://github.com/jlucasmcrell/ComfyUI-H3-Multishot">jlucasmcrell</a></li></ul>

v1.2

Update: Added I2V (Image to Video) Simple Version!- Perfect for single-image animation with original MP3 lip-sync.- Cleaner and lighter node setup compared to R2V.- Same zero audio corruption / perfect Japanese vocal sync.

MiniMax H3 - Original Audio Lip-Sync Workflow