Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Converts natural language prompts into Strudel live-coding patterns, renders them offline to WAV and MP3 without a browser or DAW, and streams the result to Discord voice channels via OpenClaw. Also runs a full audio deconstruction pipeline: separate an MP3 into stems with Demucs, extract samples with librosa, and build a generative composition from the extracted material.
Produces finished audio files from a text prompt in a fully automated pipeline without requiring a browser, DAW, or manual audio editing.
When to use it
- Generating background music for a game from a mood and tempo description
- Extracting drum and bass stems from an existing track to use as samples
- Streaming a live-coded ambient set to a Discord voice channel
- Building a custom sample library from your own recorded audio
- Rendering scored music for video or podcast production without a DAW
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: /strudel dark ambient tension, 65bpm
- 1Parse mood and tempo from the prompt using the mood-parameters reference
- 2Write a Strudel .js composition file with phrygian key, sparse drones, and a low filter cutoff
- 3Spawn a background sub-agent to run the chunked offline renderer with a 120-second timeout
- 4Convert the rendered WAV to MP3 with ffmpeg at 192k bitrate
- 5Post the MP3 as a file attachment or stream it to Discord VC via vc-play.mjs
An MP3 of a dark ambient track at 65 BPM, posted to Discord or saved as a local file