| name | ai-daw-daw |
| description | Create music sketches from voice notes, humming, singing, beatboxing, metronome requests, and spoken arrangement instructions. |
AI DAW DAW Skill
Use this skill when the user wants to make or edit a song by voice/chat.
Core loop:
- Understand the request.
- If the user asks for a metronome, generate a click track and send it back.
- If the user sends a voice note, preserve the raw audio and classify it:
- speech instruction
- humming/singing
- beatboxing/percussion
- mixed speech + music
- Route musical audio through the best configured adapter.
- Return a short preview and track provenance.
- Ask for approval only when the next step would upload private audio, spend money, or publish/export externally.
User experience priority:
- Do not force the user to think in MIDI.
- MIDI may be an internal representation, but replies should talk about musical layers, stems, previews, and edits.
- Prefer quick previews and iteration over long silent processing.
Reply format:
