---

name: nepa-audio-workspace description: AI-powered audio studio workspace for the NEPA AI platform. Handles audio trimming, effects, stem separation (demucs), music generation, and mixing. Use when editing audio files, separating stems, or generating music for NEPA AI projects.

Audio Studio Workspace

836 lines — 33 methods. Professional AI audio editing + music production.

Prerequisites

pip install pydub torch transformers demucs openai numpy scipy soundfile librosa ffmpeg-python whisper-openai
sudo apt install ffmpeg

C++ acceleration: libaudio.a (NEPA Omni Audio DSP) — included in the AXON product ZIP.

Source

/home/billk/projects/nepa-ai-monorepo.BAK/vscode_forked/nepa-ai-backend/audio_studio_workspace.py

Setup

cd ~/projects/nepa-ai-monorepo.BAK/vscode_forked/nepa-ai-backend
python run_nepa.py

Key Methods

| Method | Description | |--------|-------------| | trim_audio(path, start, end) | Trim to time range (pydub) | | separate_stems(path) | Demucs stem separation (vocals/drums/bass/other) | | generate_music(prompt, duration) | MusicGen / AudioCraft generation | | remove_silence(path, threshold) | Remove silence gaps | | normalize_audio(path) | Loudness normalization | | apply_effect(path, effect) | Reverb, EQ, compression, etc. | | transcribe_audio(path) | Whisper transcription | | mix_tracks(tracks, levels) | Multi-track mixing | | master_audio(path) | AI mastering pipeline | | convert_format(path, fmt) | Convert to mp3/wav/flac/aac | | restore_audio(path) | Audio restoration (noise reduction) | | generate_sfx(prompt) | Sound effect generation |

Example Usage

import asyncio
from audio_studio_workspace import AudioStudioWorkspace

async def main():
    ws = AudioStudioWorkspace()
    
    # Separate stems from a BMX video track
    stems = await ws.separate_stems("background_music.mp3")
    # Returns paths to: vocals.wav, drums.wav, bass.wav, other.wav
    
    # Generate background music
    music = await ws.generate_music("lofi hip hop, chill, 120bpm, 30 seconds")
    print(f"Generated: {music['output_path']}")
    
    # Trim and normalize for social media
    trimmed = await ws.trim_audio("podcast.mp3", start=10.0, end=60.0)
    normalized = await ws.normalize_audio(trimmed["output_path"])

asyncio.run(main())

API Endpoints

POST /api/audio/upload
POST /api/audio/trim
POST /api/audio/separate-stems
POST /api/audio/generate-music
POST /api/audio/transcribe
POST /api/audio/normalize
POST /api/audio/convert
POST /api/audio/mix
POST /api/audio/master

AXON Store

Available at: https://axon.nepa-ai.com (audio-workspace, $67) Download: /downloads/audio-workspace.zip

---

Full version available at https://axon.nepa-ai.com/products

This skill is part of the NEPA AI Axon workspace suite. Buy once, use forever. No subscription required.