Wraps the RunComfy CLI to generate video with HappyHorse 1.0, currently ranked #1 on Artificial Analysis Video Arena. The model outputs native 1080p with synchronized audio in the same pass, supports multi-shot character consistency, and accepts prompts in six languages. Duration caps at 15 seconds, five aspect ratios only, and the text-to-video endpoint doesn't take image conditioning. If you need lip-synced dialogue from external audio or ultra-fast iteration, the skill documentation steers you to Seedance 2.0 or LTX 2 instead. Most useful when you want broadcast-ready video with in-scene voiceover or ambient sound without a second audio pipeline, or when blind-vote leaderboard rank matters for your use case.
npx skills add https://github.com/runcomfy-com/skills --skill happyhorse-1-0