ace-step

ACE Step is a RunComfy-powered AI music generation skill that enables users to create, modify, and extend music through the RunComfy CLI. It provides access to StepFun-AI's ACE Step open-weights music foundation models, helping users generate music from descriptive tags, genres, moods, instruments, and optional lyrics. In addition to creating new tracks from text prompts, the skill supports editing existing audio through inpainting and outpainting workflows, making it useful for both music creation and targeted audio refinement without requiring a complete re-render of a track.

The skill exposes four CLI-accessible endpoints: ACE Step text-to-audio, ACE Step 1.5 text-to-audio, audio-inpaint, and audio-outpaint. The base ACE Step model focuses on cost-efficient, tag-driven music generation with stereo output ranging from 5 seconds to 4 minutes. ACE Step 1.5 adds support for lyrics in more than 50 languages and improved handling of structured lyrics with section markers. Audio inpainting allows regeneration of a specified time range within an existing track, while audio outpainting extends tracks by adding content before, after, or around existing audio. The skill is designed to work through the RunComfy CLI and requires authentication through a RunComfy account or token.

This skill is suitable for musicians, content creators, game developers, audio producers, and AI workflow builders who need flexible music generation capabilities. Common use cases include creating background music, jingles, game loops, multilingual vocal tracks, replacing problematic song sections, refining choruses or bridges, and extending existing tracks with new intros or outros. It is particularly valuable for users seeking a cost-effective music generation workflow that integrates into command-line and automated production environments.