Generate videos from text prompts or images using Pruna's optimized models through the inference.sh CLI. The skill wraps three models: P-Video for high-quality text/image-to-video with audio support, WAN-T2V for fast text-to-video, and WAN-I2V for animating still images. All models output 720p or 1080p video with Pruna's speed optimizations. Good for prototyping video content, creating social media clips, or animating product mockups without running local GPU inference. The WAN models are particularly economical at $0.05-0.11 per video, while P-Video offers more features like audio sync and higher resolution options.
npx skills add https://github.com/inferen-sh/skills --skill p-video