hunyuan-video
https://wavespeed.ai/ Context parallel attention that accelerates DiT model inference with dynamic caching
Multi-scene AI movie maker — point at any OpenAI-compatible LLM + video API, get a stitched mp4.