Voicevox MCP Server

This server enables the use of VOICEVOX-compatible speech synthesis servers (AivisSpeech / VOICEVOX / COEIROINK) via the MCP (Model Context Protocol). It can be used for speech synthesis in agent mode with Claude 3.7 using Cursor, etc.

Prerequisites

Windows Environment

• Node.js 18 or higher
• VOICEVOX ENGINE, etc. (running locally at http://localhost:50000, etc.)
• VLC Media Player (ensure it is in your system's PATH)

Docker Environment (WSL2)

• Docker and Docker Compose
• WSL2
• VOICEVOX ENGINE, etc. (running locally or in Docker)
• A Linux environment where sudo apt install libsdl2-dev pulseaudio-utils pulseaudio has been executed
• Access permissions to /mnt/wslg

Installation and Configuration

1. Clone the repository bash git clone https://github.com/Dosugamea/voicevox-mcp-server.git cd voicevox-mcp-server

2. Install dependencies bash npm install

3. Set up environment variables Copy .env_example to create a .env file and modify the settings as needed: plaintext VOICEVOX_API_URL=http://localhost:50021 VOICEVOX_SPEAKER_ID=1

How to Run

Running on Windows Environment

Please start the server separately from the editor following the steps below. bash npm run build npm start

Running in Docker Environment

No separate action from the editor is required. Since it starts in stdio mode, it cannot be run directly.

Configuration

For Running on Windows Environment

Add the following to mcp.json. If the connection is unstable and gets disconnected, please reconnect. json "voicevox": { "url": "http://localhost:10100/sse" }

For Running in Docker Environment

Add the following to mcp.json. (Operation in the author's environment has not been verified.) json { "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" } } } }

About Speaker IDs

The speaker ID varies depending on the VOICEVOX model being used. By default, it uses "1" (Shikoku Metan). If you want to use another speaker ID, change the VOICEVOX_SPEAKER_ID environment variable.

You can check the list of speaker IDs through the /speakers endpoint of the VOICEVOX ENGINE API. Example: curl http://localhost:50021/speakers## Troubleshooting

• Connection Error with VOICEVOX: Please ensure that the VOICEVOX ENGINE is running and that the API URL is correctly set.
• No Audio Playback: Verify that VLC is properly installed and that the path is correctly configured.
• Audio Output Issues in Docker Environment: Check if the pulseaudio settings are correct.

Information for Developers

• If you wish to contribute to the source code, please create an Issue or submit a Pull Request.
• For bug reports or feature requests, please use the Issue feature on GitHub.

License

MIT License