2.6k
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill openclaw/skills --skill acestep-songwriting- _meta.json292 B
- SKILL.md7.7 KB
Overview
This skill is a professional songwriting guide for ACE-Step that helps you craft precise captions, structured lyrics, and appropriate music parameters before generating a track. It focuses on caption design, lyric structure tags, and metadata like BPM, key, time signature, and duration. Use it to plan songs with clarity so the generation model produces consistent, singable results.
How this skill works
The guide inspects the desired style across dimensions: genre, emotion, instruments, timbre, era, and vocal characteristics, and turns those into a concise caption that anchors generation. It provides structured lyric templates with section tags, vocal hints, and energy markers so lyrics align with arrangement and instrumentation. It also helps calculate duration and recommends when to set BPM, key, and time signature versus letting the model infer them.
When to use it
- Before generating a new song in ACE-Step to set a clear creative direction
- When you need a precise caption that guides instrumentation and production style
- To convert a song idea into singable, structured lyrics with section tags
- When matching tempo, key, or duration to existing material
- When you want consistent output across multiple generation runs
Best practices
- Make captions specific: combine style + emotion + instruments + timbre (e.g., "warm 80s synthwave, melancholic, female breathy vocal").
- Keep lyrics lines short (6–10 syllables) and separate sections with blank lines for clarity.
- Use dedicated parameters for BPM, key, and duration instead of embedding them in the caption.
- Align caption instruments with lyric tags (e.g., caption includes piano → lyrics include [Piano Interlude]).
- Avoid conflicting descriptors; if you want transitions, state them explicitly (start soft strings, become metal).
Example use cases
- Create a caption and lyrics for a 3-minute pop ballad with piano and strings before generation.
- Plan a mid-tempo R&B song: caption for mood and instruments, lyrics with [Pre-Chorus] and [Chorus] tags and vocal hints.
- Prepare metadata to match an existing beat by setting BPM and duration for synchronization.
- Design a dynamic EDM track: caption for synth texture and a lyric script with [Build] and [Drop] sections.
- Write singable folk verses with a single coherent metaphor and matching [Guitar Solo] tag.
FAQ
No. Use the dedicated parameters (--bpm, --key, --time-signature) so the model treats musical metadata separately from the caption description.
How long should a typical song be?
Use the duration guidelines: 2 verses + 2 choruses ~120–150s; add a bridge or intro/outro to reach 180–240s depending on BPM and structure.