MCP Tts Voicevox

🚀 MCP TTS Voice Synthesis Server Using VOICEVOX

This project provides a voice synthesis server using VOICEVOX, featuring efficient queue management and pre - fetching for smooth voice playback.

🚀 Quick Start

Prerequisites

• Node.js
• VOICEVOX Engine

Installation

npm install -g @kajidog/mcp-tts-voicevox

Usage as an MCP Server

1. Start the VOICEVOX Engine.
2. Start the MCP Server.

npx @kajidog/mcp-tts-voicevox

Direct Import in the Project

npm install @kajidog/mcp-tts-voicevox

import { VoicevoxClient } from "@kajidog/mcp-tts-voicevox";

// Initialize the client
const client = new VoicevoxClient({
  url: "http://localhost:50021", // URL of the VOICEVOX Engine
  defaultSpeaker: 1,             // Default speaker ID (optional)
  defaultSpeedScale: 1.0         // Default speed (optional)
});

// Convert text to speech and play
await client.speak("你好");

// Generate an audio file from text
const filePath = await client.generateAudioFile("你好", "./output.wav");

// Clear the queue
await client.clearQueue();

✨ Features

• Queue Management Function - Efficiently handle multiple voice synthesis requests.
• Pre - fetching - Generate the next voice in advance for smooth playback.

💻 Usage Examples

Basic Usage

import { VoicevoxClient } from "@kajidog/mcp-tts-voicevox";

// Initialize the client
const client = new VoicevoxClient({
  url: "http://localhost:50021", // URL of the VOICEVOX Engine
  defaultSpeaker: 1,             // Default speaker ID (optional)
  defaultSpeedScale: 1.0         // Default speed (optional)
});

// Convert text to speech and play
await client.speak("你好");

Advanced Usage

// Generate an audio file from text
const filePath = await client.generateAudioFile("你好", "./output.wav");

// Clear the queue
await client.clearQueue();

📚 Documentation

Main Functions

• Text Reading (speak) - Convert text to speech and play.
• Query Generation (generate_query) - Create a query for voice synthesis.
• File Generation (synthesize_file) - Generate a voice file from a query.
• Queue Cleaning (clear_queue) - Clear all current voice synthesis tasks.

Environment Variables

📄 License

ISC