VOICEVOX

🚀 Voicevox MCP Server

An interoperable speech synthesis server that leverages the speech synthesis technologies of VOICEVOX, AivisSpeech, and COEIROINK through the Model Context Protocol (MCP). It is suitable for speech synthesis in the proxy mode of Claude 3.7 in tools like Cursor.

🚀 Quick Start

Prerequisites

Windows Environment

• Node.js 18 or higher
• VOICEVOX ENGINE (running locally, e.g., at http://localhost:50000)
• VLC Media Player (with the path configured)

Docker Environment (WSL2)

• Docker and Docker Compose
• WSL2
• VOICEVOX ENGINE (running locally or in Docker)
• A Linux environment with sudo apt install libsdl2-dev pulseaudio-utils pulseaudio installed
• Access to /mnt/wslg

📦 Installation

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 and rename the .env_example file to .env, and modify the settings as needed:

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

Execution Methods

Windows Environment

Execute the following commands in the editor to start the server:

npm run build
npm start

Docker Environment

No additional actions are required in the editor. Since it is in stdio mode, it cannot be run directly.

Configuration Methods

Windows Environment

Add the following content to mcp.json: To avoid disconnections due to unstable connections, ensure automatic reconnection.

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

Docker Environment

Add the following content to mcp.json: (The author's environment has not confirmed whether it works properly.)

{
    "tools": {
        "voicevox": {
            "command": "cmd",
            "args": [
                "/c",
                "docker",
                "run",
                "-i",
                "--rm",
                "-v",
                "/mnt/wslg:/mnt/wslg",
                "-e",
                "PULSE_SERVER",
                "-e",
                "SDL_AUDIODRIVER",
                "-e",
                "VOICEVOX_API_URL",
                "-e",
                "VOICEVOX_SPEAKER_ID",
                "your-local-docker-image-name"
            ],
            "env": {
                "PULSE_SERVER": "unix:/mnt/wslg/PulseServer",
                "SDL_AUDIODRIVER": "pulseaudio",
                "VOICEVOX_API_URL": "http://host.docker.internal:50031",
                "VOICEVOX_SPEAKER_ID": "919692871"
            }
        }
    }
}

Speaker ID

The speaker ID varies depending on the VOICEVOX model used. By default, "1" (Shikoku Mikan) is used. If you need to use a different speaker ID, modify the environment variable VOICEVOX_SPEAKER_ID.

You can confirm the list of speaker IDs through the /speakers endpoint of the VOICEVOX ENGINE API. For example: curl http://localhost:50021/speakers

Troubleshooting

• VOICEVOX connection error: Confirm that the VOICEVOX ENGINE is started and the API URL is correctly configured.
• Audio not playing: Confirm that VLC is correctly installed and the path is configured.
• Audio output issues in the Docker environment: Check the pulseaudio settings.

Developer Information

• If you want to contribute code, please create an Issue or submit a Pull Request.
• For error reports or feature requests, please use the GitHub Issue function.

📄 License

MIT License