acestep

Generates music, covers, and audio stems with ACE-Step 1.5 on RunPod.

Updated Jul 16, 2026
One-click install
npx skills add https://github.com/X-manist/Cohmira --skill acestep-x-manist
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: acestep
Source: https://github.com/X-manist/Cohmira/tree/main/src/builtin-plugins/openmontage/.agents/skills/acestep
Command: npx skills add https://github.com/X-manist/Cohmira --skill acestep-x-manist

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Producing background music, jingles, vocal tracks, and isolated instrument stems for video production normally requires licensing libraries or hiring composers. This Skill generates original music from text prompts, lyrics, and scene presets using the open-source ACE-Step 1.5 model running on RunPod serverless. ## Core Features & Use Cases - Text-to-Music Generation: Create instrumental tracks or full songs with lyrics, controlling BPM, key, duration, and seed for reproducible results. - Scene Presets for Video: Use built-in presets like corporate-bg, tension, hopeful, cta, and lofi matched to common video scene types with precise durations. - Cover and Stem Extraction: Apply style transfer to reference audio with adjustable cover strength, or isolate vocals, drums, bass, and other stems from mixed audio. - Use Case: While editing a product demo video, generate a 20-second tension track for the problem scene, a 60-second lofi track under the screen recording, and a 15-second branded CTA jingle, all mixed under voiceover at 10-20% volume. ## Quick Start Generate a 60-second upbeat corporate background track at 110 BPM in C Major and save it as bg.mp3 for my presentation video.

Frequently Asked Questions about acestep

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate background music for a video with AI?

Run music_gen.py with a text prompt describing genre, mood, and instruments, plus a duration matching your scene length. Use scene presets like corporate-bg, lofi, or tension to quickly match common video scene types.

How to generate a song with lyrics using ACE-Step?

Pass lyrics with the --lyrics flag using structure tags like [Verse] and [Chorus] to control song sections. Keep 6-10 syllables per line, use UPPERCASE for high vocal intensity, and parentheses for background vocals.

What do I need to run ACE-Step music generation?

You need a RunPod account with RUNPOD_API_KEY and RUNPOD_ACESTEP_ENDPOINT_ID set in a .env file. Run the tool with --setup to create the serverless endpoint before generating music.

Can ACE-Step extract vocals or stems from a song?

Yes, the extract task isolates individual tracks including vocals, drums, bass, guitar, piano, strings, and more from mixed audio. For video files, extract the audio with FFmpeg first, then run stem separation.

When should I not use ACE-Step for audio tasks?

ACE-Step is not suited for voice cloning, speech narration, or sound effects. Use Qwen3-TTS or ElevenLabs for voice work and ElevenLabs SFX for sound effects instead.