VoiceBox

Provides an MCP interface to VoiceBox TTS, exposing speak, status, speakers, health, and metrics endpoints.
  • typescript

0

GitHub Stars

typescript

Language

5 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "shunsukehayashi-voicebox-mcp": {
      "command": "node",
      "args": [
        "/Users/shunsukehayashi/dev/voicebox-mcp/dist/index.js"
      ],
      "env": {
        "VOICEBOX_API_URL": "http://localhost:5001"
      }
    }
  }
}

You deploy the VoiceBox MCP Server to bridge VoiceBox TTS with VOICEVOX through the Model Context Protocol (MCP). It exposes endpoints to convert text to speech, check task status, list speakers, monitor health, and retrieve metrics, enabling seamless integration with MCP clients and call flows.

How to use

Use the MCP client to interact with the VoiceBox MCP Server for common actions like creating a speak task, polling its progress, and listing available speakers. Typical workflows include converting text to speech with a selected speaker, checking on long-running tasks, and verifying system health before starting production workloads.

How to install

Prerequisites: you need Node.js and npm installed on your machine.

Install and build the MCP server locally.

Notes: you should also have VoiceBox TTS API running and VOICEVOX API server enabled (port 50021) before starting the MCP server.

Configuration and usage notes

To configure the MCP server connection to VoiceBox, you will run the MCP server locally and provide the VoiceBox API URL via environment variables. You can also configure a client like Claude to use this MCP server by pointing it to the appropriate command and environment.

Additional details

The MCP server exposes these functions: voicebox_speak to convert text to speech, voicebox_status to check task status, voicebox_speakers to list available speakers, voicebox_health to check system health, and voicebox_metrics to get system metrics. You can use these functions to build robust, interactive TTS workflows in your applications.

Claude code configuration example

{
  "mcpServers": {
    "voicebox": {
      "command": "node",
      "args": ["/Users/shunsukehayashi/dev/voicebox-mcp/dist/index.js"],
      "env": {
        "VOICEBOX_API_URL": "http://localhost:5001"
      }
    }
  }
}

Available Speakers

The VoiceBox MCP Server exposes a set of predefined voices you can select when issuing a speak request. The following IDs map to speakers: 0 四国めたん (あまあま), 1 四国めたん (ノーマル), 2 四国めたん (セクシー), 3 ずんだもん (ノーマル), 4 ずんだもん (あまあま), 5 ずんだもん (悲嘆), 6 春日部つむぎ (ノーマル), 7 春日部つむぎ (あまあま), 8 春日部つむぎ (ツンデレ).

Development

Development commands you may run locally while building and testing the MCP server include:
- `npm run dev` to watch for changes
- `npm run build` to build the server
- `npm start` to run the server

Available tools

voicebox_speak

Convert text to speech for a given speaker and returns an audio stream or file reference.

voicebox_status

Check the status of an ongoing speak task.

voicebox_speakers

List all available VOICEVOX speakers that can be used for TTS.

voicebox_health

Return current health indicators for the MCP server and its dependencies.

voicebox_metrics

Provide system metrics such as CPU, memory, and request throughput.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational