Tag-driven music generation, inpainting, and outpainting via StepFun-AI's ACE Step model.

• Four endpoints: text-to-audio (base and 1.5 variants with 50+ language support), audio-inpaint (regenerate time ranges), and audio-outpaint (extend before/after)
• Pricing $0.0002–0.0003 per second of audio, roughly 27× cheaper than ElevenLabs Music; supports 5–240 second generation per call
• Composition driven by comma-separated tags (genre, mood, instruments, BPM) with optional structured lyrics using section markers ([Verse], [Chorus], [Bridge])
• Invoked via runcomfy CLI; requires authentication and supports seed-based reproducibility