MCP TTS VOICEVOX

MCP TTS VOICEVOX is an MCP server that integrates an external VOICEVOX text-to-speech engine with LLM interfaces like Claude Desktop and ChatGPT. It targets developers, content creators, and AI enthusiasts seeking expressive Japanese speech synthesis directly inside conversation threads. By bridging the Model Context Protocol with the VOICEVOX Engine API, the server enables voice generation across dozens of distinct character personas. Users can execute standard audio synthesis via tools like voicevox_speak, which supports queuing, speaker switching, inline accent phrase customization, and file exports. Additionally, the server features an interactive MCP Apps UI audio player that allows client-side playback without requiring local server audio hardware. This embedded player supports track reordering, pitch and mora adjustments, custom user dictionary management, and segmented multi-speaker dialogues, allowing language models to act as interactive vocal assistants or audio production editors.

Category: Design, Media & Creative

Tags: audio, synthesis, text to speech, voicevox

Visit MCP TTS VOICEVOX

How to install and configure MCP TTS VOICEVOX

1. Ensure Node.js 20.0.0 or higher (or Bun) is installed, and launch the external VOICEVOX Engine on your machine. 2. Optional: Install ffplay via your operating system package manager to enable low-latency streaming playback. 3. Open your Claude Desktop configuration file: - Windows: %APPDATA%\Claude\claude_desktop_config.json - macOS: ~/Library/Application Support/Claude/claude_desktop_config.json 4. Add the server entry under mcpServers: json { "mcpServers": { "tts-mcp": { "command": "npx", "args": ["-y", "@kajidog/mcp-tts-voicevox"] } } } 5. Restart Claude Desktop and ask the assistant to speak.

What you can do with MCP TTS VOICEVOX

• Triggering direct Japanese voice playback from Claude Desktop using multiple VOICEVOX character IDs within a single message. - Fine-tuning pronunciation and pitch drops by providing inline accent notation brackets during real-time speech generation. - Running an embedded MCP Apps audio player inside chat sessions to preview, adjust speed, and reorder dialogue segments. - Exporting generated multi-speaker dialogue tracks directly to numbered WAV files on the local filesystem. - Managing VOICEVOX custom user dictionary words and phonetic pronunciations directly through conversational AI prompts.

Key facts

• Open Source
• https://github.com/kajidog/mcp-tts-voicevox
• Design, Media & Creative
• audio, synthesis, text to speech, voicevox

Part of MCP Servers

What is MCP TTS VOICEVOX?

MCP TTS VOICEVOX is an open-source text-to-speech server that connects AI assistants to the VOICEVOX speech synthesis engine via the Model Context Protocol. It allows clients like Claude Desktop to speak text using various character voices, stream audio, adjust phonetic accents, and interact with an embedded audio player.

Which MCP clients work with MCP TTS VOICEVOX?

MCP TTS VOICEVOX works with Claude Desktop via stdio or mcp-remote. It also supports remote HTTP connections for ChatGPT and other hosts compatible with Model Context Protocol and MCP Apps extensions for client-side audio rendering.

How do I install MCP TTS VOICEVOX?

You can run the server directly using npx or bunx targeting @kajidog/mcp-tts-voicevox inside your Claude Desktop configuration file. Alternatively, you can deploy both the MCP server and VOICEVOX Engine together using Docker Compose with mcp-remote.

Does MCP TTS VOICEVOX require local audio hardware?

Not necessarily. Standard server-side playback via voicevox_speak requires local audio hardware, but the voicevox_speak_player tool streams audio directly into the client application UI, allowing audio playback across remote environments without server-side sound devices.

Is MCP TTS VOICEVOX open source?

Yes, MCP TTS VOICEVOX is an open-source project hosted on GitHub under the repository kajidog/mcp-tts-voicevox. Anyone can inspect the code, customize integration scripts, or run containerized instances freely.