Pruna AI's multimodal video model built for speed and iteration — supports text, image, and audio-to-video in a unified endpoint. 1-10s, up to 1080p.