A timeline editor for MiniMax H3 inside ComfyUI - storyboard prompts, first/last keyframes, image/video/audio references, joint audio, live sampling preview, retakes and shot chaining.
Python SDK for the MiniMax H3 API — text-to-video, image-to-video, and first/last-frame video generation via Muapi.
🎬 Curated MiniMax H3 video generation prompts — cinematic, ads, anime, UGC, product videos, and more. Includes playable examples and creator attribution.
NVIDIA Sol-Attn for ComfyUI / Triton kernel on SM89/SM90/SM100/SM120, with zero-copy MiniMax H3 nodes: memory-efficient attention, scheduled tau with graph preview, and feed-forward chunking. Measured 1.14–1.44× vs SageAttention and −37% MLP peak VRAM on H3
Timeline-based video editor for ComfyUI with scenes, project assets, gallery, images as keyframes, render staging, and takes for iterative long-form generation.
AI驱动的影视广告创作平台/AI-driven platform for film and television advertising creation
Measured SM121 compatibility recipe for MiniMax H3 FL2VA on one NVIDIA DGX Spark with vLLM-Omni and online FP8.
Experimental, entirely AI-coded ComfyUI nodes for MiniMax H3 T2I, I2I, reference editing, arbitrary frames, and optimized still selection.
MiniMax H3 FirstBlockCache and low-step preview nodes for ComfyUI | 神都猫
Timestep-cache acceleration for MiniMax H3 in ComfyUI — open source, cross-platform, ~3x speedup measured
🎬 A Codex-first MiniMax H3 video production plugin: 👤 character & 🏙️ scene design, 🧩 storyboards, 🎞️ ComfyUI generation, ✂️ post-production, and ✅ QC.
NVIDIA's Sol-Attn training-free sparse attention as an opt-in ComfyUI model patch - including consumer Blackwell (SM120 / RTX 50-series), which upstream gates off.
The ultimate MiniMax H3 prompt library for AI video generation—featuring source-verifiable prompts, official input assets, playable outputs, and resources from MiniMax’s official guide and X posts.
MiniMax H3 (Hailuo 3.0) prompt library — 135 prompts, each with a clip you can watch. Browse at apimodels.app/minimax-h3-prompts
Original outfit prompt director plus EvoLink MiniMax H3 and Seedance 2 image-to-video Skills
A Claude Code skill for MiniMax H3 — the open-weight omni-modal video model. Prompt framework, ComfyUI model/quant reference and troubleshooting for T2V, I2V, FLF2V and R2V (reference-to-video).
Original MiniMax H3 prompt library, MuAPI Python client, API examples, and current model notes.
Curated MiniMax H3 / Hailuo 3.0 AI video prompts, examples, templates, and browser-first X discovery workflow.
Pure MLX MiniMax-H3 text-to-video-and-audio inference for Apple Silicon.
ComfyUI custom nodes for MiniMax H3 text-to-video, image-to-video, and multimodal reference-to-video generation through Muapi.
Add a description, image, and links to the minimax-h3 topic page so that developers can more easily learn about it.
To associate your repository with the minimax-h3 topic, visit your repo's landing page and select "manage topics."