Ai Video Generation

Gives you command-line access to 40+ AI video models through inference.sh, including Google Veo 3.1, ByteDance's Seedance 2.0, and Alibaba's HappyHorse. You can do text-to-video, animate static images, create talking avatars with lipsync, edit existing videos with natural language prompts, and add foley sound effects. The range is legitimately impressive, from quick social media clips with Veo 3.1 Fast to character-preserving animations with HappyHorse's reference-to-video mode. Most useful if you're building content pipelines or need to try multiple models without managing separate APIs. Fair warning: you'll need the belt CLI installed first, and costs vary wildly between models.

Install to Claude Code

npx -y skills add inference-sh-skills/skills --skill ai-video-generation --agent claude-code