Audio Player MCP Bundle for Claude Desktop with ElevenLabs integration. Generate speech, sound effects, and music, or play any local audio file.

Features

• Text-to-Speech - Generate speech from text using ElevenLabs voices
• Sound Effects - Create sound effects from text descriptions
• Music Generation - Compose music from prompts
• Audio Playback - Play any local audio file with a built-in player UI
• Playlist Support - Queue multiple tracks with playlist view
• Playback Controls - Progress bar, speed adjustment, and standard controls

Installation

Build the MCPB Bundle

# Clone the repository
git clone https://github.com/elevenlabs/elevenlabs-mcp-player.git
cd elevenlabs-mcp-player

# Install dependencies
npm install

# Build and pack the bundle
npm run pack

This creates elevenlabs-player.mcpb in the project root.

Install in Claude Desktop

Open the bundle with Claude Desktop:

open elevenlabs-player.mcpb

Or double-click the .mcpb file in Finder.

Configuration

To use the ElevenLabs generation features (TTS, sound effects, music), configure your API key in Claude Desktop:

1. Open Claude Desktop settings
2. Navigate to the ElevenLabs Player extension settings
3. Enter your ELEVENLABS_API_KEY

Environment Variable	Description	Default
ELEVENLABS_API_KEY	Your ElevenLabs API key (required for generation)	-
ELEVENLABS_OUTPUT_DIR	Directory to save generated audio	Desktop

Usage

Generate Speech

"Say 'Hello, world!' using ElevenLabs"

"Generate speech for this text: Welcome to the future of AI"

Generate Sound Effects

"Create a sound effect of thunder and rain"

"Generate the sound of a spaceship taking off"

Generate Music

"Compose a calm piano melody for relaxation"

"Generate upbeat electronic music for a workout"

Play Local Audio

"Play the audio file at /Users/me/Music/song.mp3"

"Play these audio files: /path/to/track1.mp3 and /path/to/track2.mp3"

Tools

generate_tts

Generates speech from text using ElevenLabs text-to-speech.

Parameter	Required	Description
text	Yes	The text to convert to speech
voice_id	No	ElevenLabs voice ID (default: Juniper)
model_id	No	Model ID (default: eleven_v3)
title	No	Display title for the track

generate_sound_effect

Generates a sound effect from a text description.

Parameter	Required	Description
prompt	Yes	Description of the sound effect
duration_seconds	No	Duration in seconds
title	No	Display title for the track

generate_music

Generates music from a text prompt.

Parameter	Required	Description
prompt	Yes	Description of the music (genre, mood, instruments)
duration_seconds	No	Duration in seconds
instrumental	No	Force instrumental only (no vocals)
title	No	Display title for the track

play_audio

Plays one or more local audio files.

Parameter	Required	Description
tracks	Yes	Array of track objects
tracks[].filePath	Yes	Absolute path to the audio file
tracks[].title	Yes	Display title
tracks[].artist	No	Artist name

Supported Audio Formats

• MP3 (.mp3)
• WAV (.wav)
• OGG (.ogg)
• M4A (.m4a)
• AAC (.aac)

Development

Requirements

• Node.js 20+
• npm

Testing Locally

Use the MCP Inspector to test the server without Claude Desktop:

# Install dependencies
npm install

# Build and run with MCP Inspector
npm run dev

This opens a browser UI at http://localhost:6274 where you can:

• View available tools
• Test tool calls with sample inputs
• Inspect responses and debug issues

Scripts

Command	Description
npm run dev	Build and test with MCP Inspector
npm run build	Build UI and server
npm run pack	Create MCPB bundle

Project Structure

├── manifest.json          # MCPB bundle manifest
├── server.ts              # MCP server with tool registration
├── src/
│   ├── mcp-app.tsx        # React audio player UI
│   └── components/ui/     # UI components
├── dist/
│   ├── server.js          # Compiled server
│   └── mcp-app.html       # Bundled UI
└── .mcpbignore            # Files to exclude from bundle

Privacy

This extension runs locally on your machine. Local audio files are read from your filesystem and are not transmitted externally.

When using ElevenLabs generation features (TTS, sound effects, music), your text prompts are sent to the ElevenLabs API. Generated audio is saved locally to your configured output directory.

For more information about ElevenLabs' data practices, see the ElevenLabs Privacy Policy.

License

MIT License - see LICENSE for details.