Voicevox MCP Server

🚀 Voicevox MCP Server

This is a speech synthesis server that uses Voicevox, AivisSpeech, or COEIROINK through the Model Context Protocol (MCP). It is suitable for speech synthesis in proxy mode when using Claude 3.7 in tools like Cursor.

🚀 Quick Start

Prerequisites

Windows Environment

• Node.js version 18 or higher.
• Voicevox ENGINE installed and running (locally accessible via http://localhost:50000).
• VLC Media Player added to the system path.

Docker Environment (WSL2)

• Docker and Docker Compose installed.
• WSL2 environment.
• Voicevox ENGINE installed and running (locally or inside Docker).
• sudo apt install libsdl2-dev pulseaudio-utils pulseaudio installed on the Linux system.
• Access to /mnt/wslg.

Installation and Configuration

1. Clone the repository

git clone https://github.com/Dosugamea/voicevox-mcp-server.git
cd voicevox-mcp-server

2. Install dependencies

npm install

3. Configure environment variables Copy .env_example to a .env file and modify it as needed:

VOICEVOX_API_URL=http://localhost:50021
VOICEVOX_SPEAKER_ID=1

Running the Server

Windows Environment

Execute the following commands in the editor and other windows respectively to start the server:

npm run build
npm start

Docker Environment

No need to operate in the editor, just run it directly. Since it is in stdio mode, it cannot be started directly.

Configuration

Windows Environment

Add the following content to mcp.json:

        "voicevox": {
            "url": "http://localhost:10100/sse"
        }

If the connection is unstable, consider reconnecting.

Docker Environment

Add the following content to mcp.json (not verified in the author's environment):

{
    "tools": {
        "voicevox": {
            "command": "cmd",
            "args": [
                "/c",
                "docker",
                "run",
                "-i",
                "--rm",
                "-v",
                "/mnt/wslg:/mnt/wslg",
                "-e",
                "PULSE_SERVER=unix:/tmp/pulseaudio.sock",
                "-e",
                "QT_QPA_PLATFORM=offscreen",
                "dosugamea/voicevox",
                "python3",
                "main.py"
            ],
            "env": {
                "LD_LIBRARY_PATH": "/usr/lib:/$LD_LIBRARY_PATH",
                "PULSE_SERVER": "unix:/tmp/pulseaudio.sock",
                "QT_QPA_PLATFORM": "offscreen"
            }
        }
    }
}

About Speaker ID

• VOICEVOX_SPEAKER_ID=1 corresponds to the "Yuki" character.
• For other IDs and their corresponding characters, refer to https://voicevox.ai/speaker/.

Troubleshooting

Connection Errors with Voicevox

• Ensure that Voicevox is correctly installed and running. In particular, the Wails application version should be v3.0.0 or higher.
• Check if other processes are occupying the relevant ports.

Packaging Failures

• Install the .NET SDK: dotnet install sdk.
• Ensure the correct Node version: Use nvm use 16.

Unable to Start Voicevox in Docker

• Ensure that Docker is correctly installed and running.
• Check if other processes are occupying the relevant ports.
• Ensure that the Wails application version is v3.0.0 or higher.

Configuration Errors

• Check the console output for specific error information.
• Run in debug mode using npm run dev to facilitate problem discovery and resolution.