AudioShake Pet

AI-powered audio stem separation and transcription plugin for OpenPets. Separate vocals, instruments, dialogue, and effects from audio/video files using AudioShake's advanced AI models.

Features

• Stem Separation: Isolate vocals, instrumental, drums, bass, piano, wind, and more
• Dialogue/Music/Effects: Separate dialogue, music, and effects for video/podcast editing
• Multi-Speaker Isolation: Separate individual speakers from multi-person audio
• Lyric Transcription: Extract and align lyrics with vocal isolation
• Multiple Output Formats: WAV, MP3, FLAC, OGG support
• Residual Audio: Optionally include everything NOT in the primary stem
• Task Management: Create, monitor, and retrieve processing tasks
• Usage Statistics: Track API usage and costs

Setup

1. Get Your API Key

1. Visit AudioShake Dashboard
2. Create an account (or sign in)
3. Navigate to the API section
4. Generate your API key

2. Configure Environment

Create a .env file in the project root:

AUDIOSHAKE_API_KEY=your_api_key_here

# Optional: Override API base URL (defaults to https://api.audioshake.ai)
# AUDIOSHAKE_HOST=https://api.audioshake.ai

3. Test Connection

cd pets/audioshake
opencode run "test audioshake connection"

Available Tools

audioshake-test-connection

Test API connectivity and configuration status.

Usage: test audioshake connection

audioshake-create-task

Create a stem separation task from a publicly accessible URL.

Parameters:

• url (required): HTTPS URL to audio/video file
• model (required): Separation model (vocals, instrumental, drums, bass, piano, wind, dialogue, effects, music_removal, multi_voice)
• formats (required): Output formats (array of wav, mp3, flac, ogg)
• residual (optional): Include residual audio (default: false)
• variant (optional): Model variant selection
• language (optional): Language code for transcription (e.g., en, es, fr)

Example:

opencode run "separate vocals from https://example.com/song.mp3 in wav format"
opencode run "create task to isolate drums and bass from https://example.com/track.wav"

audioshake-list-tasks

List processing tasks with pagination.

Parameters:

• skip (optional): Number of tasks to skip (default: 0)
• take (optional): Number of tasks to return (default: 20, max: 100)

Example:

opencode run "list my audioshake tasks"
opencode run "show first 10 audioshake tasks"

audioshake-get-task

Get detailed status and results for a specific task.

Parameters:

• taskId (required): The task ID to retrieve

Example:

opencode run "get status of task abc123"
opencode run "show results for audioshake task xyz789"

audioshake-get-statistics

Get usage statistics for your account.

Example:

opencode run "show audioshake usage statistics"
opencode run "get my audioshake account stats"

audioshake-upload-from-link

Upload an audio file from URL to create a reusable asset.

Parameters:

• link (required): HTTPS URL to the file
• name (optional): Asset name (3-100 characters)

Example:

opencode run "upload https://example.com/audio.mp3 as My Song"

Supported Models

• vocals: Isolate vocal tracks
• instrumental: Extract instrumental (non-vocal) audio
• drums: Isolate drum tracks
• bass: Extract bass lines
• piano: Isolate piano parts
• wind: Extract wind instruments
• dialogue: Separate dialogue from background
• effects: Isolate sound effects
• music_removal: Remove music, keep speech
• multi_voice: Separate individual speakers

Output Formats

• wav: Uncompressed audio (highest quality)
• mp3: Compressed audio (smaller file size)
• flac: Lossless compression
• ogg: Open-source compressed format

Example Workflows

Basic Vocal Separation

opencode run "test audioshake connection"
opencode run "separate vocals from https://example.com/song.mp3 in wav format"
opencode run "list my tasks"
opencode run "get status of task <task-id>"

Multi-Stem Separation

opencode run "create task to separate drums from https://example.com/full-mix.wav in mp3 format"
opencode run "list my recent tasks"
opencode run "show results for the first task"

With Residual Audio

opencode run "separate vocals from https://example.com/track.wav with residual in wav format"

Dialogue/Music Separation

opencode run "separate dialogue from https://example.com/podcast.mp3 in mp3 format"

API Documentation

For detailed API information, visit AudioShake Developer Docs.

Troubleshooting

"Not configured" error

• Ensure AUDIOSHAKE_API_KEY is set in your .env file
• Check that the API key is valid by testing at AudioShake Dashboard

"API error: 401" (Unauthorized)

• Your API key is invalid or expired
• Generate a new key from the AudioShake Dashboard

"API error: 400" (Bad Request)

• Check that the audio URL is publicly accessible via HTTPS
• Verify the model name and format are valid
• Ensure formats is an array (e.g., ["wav"], not just "wav")

Task stays in "pending" status

• Processing time varies by audio length and model complexity
• Check back periodically using audioshake-get-task
• Download URLs appear when status becomes "completed"

License

This plugin is part of the OpenPets ecosystem.

Support

For issues with this plugin, open an issue on the OpenPets repository.

For AudioShake API support, contact [email protected].