A ComfyUI workflow built around Capricorncd Timeline and MiniMax H3 Video Generator, with an editable timeline and examples of five video generation modes.<h3 id="features">Features</h3><ul><li>Text to Video (T2V)</li><li>Image / Multi-Image Reference to Video (I2V)</li><li>Video Reference to Video (V2V)</li><li>First &amp; Last Frame to Video (FL2V)</li><li>Audio-driven lip-sync avatars</li><li>Optional two-pass latent upscaling and refinement</li><li>Sampling previews, subtitles, audio tracks and video export</li><li>Optional face refinement and motion deblur — experimental and not extensively tested in this workflow</li></ul><h3 id="required-nodes">Required Nodes</h3><ul><li>Updated ComfyUI with native MiniMax H3 support</li><li><a rel="ugc" href="https://github.com/capricorncd/ComfyUI-Capricorncd-Timeline">ComfyUI-Capricorncd-Timeline</a></li><li>ComfyUI-Easy-Use</li><li>ComfyUI-KJNodes</li><li>Comfyui_Minimax_h3_latent_Upscaler for two-pass upscaling</li><li><a target="_blank" rel="ugc" href="https://github.com/facok/comfyui-SelfLift">facok/comfyui-SelfLift</a></li><li>FFmpeg for media processing and export</li></ul>Optional features require ComfyUI-H3-FaceRefine, ComfyUI-MAINodes (motion deblur), ComfyUI-H3-AudioRefine, ComfyUI-WanVideoWrapper (loudness normalization), or ComfyUI-H3-Motion-Context.<h3 id="models">Models</h3>Place models in the corresponding ComfyUI folders:FolderModel<code>models/diffusion_modelsMiniMax_H3-Pruned-Ref-Delta-Fused-r1024-comfy-int8-convrot.safetensorsmodels/diffusion_modelsMiniMax_H3_FL2VA_int8_convrot.safetensors</code> for FL2V<code>models/text_encodersqwen3vl_32b_heretic_minimax_h3_nvfp4.safetensors</code>, or a compatible H3 text encoder<code>models/vaeminimax_h3_video_vae_fp16.safetensors</code> and <code>minimax_h3_audio_vae_fp32.safetensorsmodels/loras</code>Matching ref2v or fl2v Turbo LoRAs for the selected 4/8-step sampling mode<code>models/latent_upscale_modelsminimax_h3_latent_upscaler_3d_fp16.safetensors</code> for two-pass upscaling<code>models/vae_approxtaeh3.safetensors</code> for RGB sampling previews<code>models/ultralytics/bboxface_yolov8m.pt</code> for face refinementSelect the appropriate model and matching LoRA before generating each mode. Motion deblur also requires a separate base model without an acceleration LoRA.<h3 id="how-to-use">How to Use</h3>Load the workflow, open the Timeline Editor, click Import ZIP in the upper-right corner, and select the corresponding project and assets ZIP package from <code>workflows/resoure</code>.Select your installed models, choose the Clips to generate, and run.<a target="_blank" rel="ugc" href="https://github.com/capricorncd/ComfyUI-Capricorncd-Timeline">Installation and project repository</a> · <a target="_blank" rel="ugc" href="https://github.com/capricorncd/ComfyUI-Capricorncd-Timeline/blob/main/docs/h3-video-generator.md">MiniMax H3 Video Generator documentation</a>

Files with SHA256: 122ddbb263692bd50a21637bcd8539f8fdec3fde78a7a7ac5fe3b7a30b30753a