gemini-tts-mcp
Gives AI assistants a voice: Gemini text-to-speech as two MCP tools.
- Domain
- Voice
- Built with
- TypeScript, Node.js, MCP SDK
- Source
- github.com/notsointresting/gemini-tts-mcp
The problem
AI command-line assistants cannot speak without someone wiring up an audio pipeline by hand.
How it works
- An MCP server over Gemini 2.5 Flash and Pro TTS.
- generate_speech and list_voices tools, with vocal style set in plain language.
- Base64 audio streaming or direct .wav files on disk.
- Packaged as a Gemini CLI extension and a Claude Desktop tool.
Where it landed
30+ voices (Aoede, Zephyr, Puck, Charon, Kore, Fenrir and more) with configurable rate limits.
Stack
- TypeScript
- Node.js
- MCP SDK
- Gemini API
- Zod