Summary
Narration you can direct, not just generate.
Most TTS tools give you a voice and a speed. Here emotion and pacing are separate dials, which is the difference between a tutorial that sounds like a tutorial and one that sounds like a phone tree.
What it is
A text-to-speech server written in TypeScript with no model downloads: ten voices, expressive controls, batch processing, and file management for what it produces.
What you get
- 10 voices, male and female, with distinct characters
- Emotion settings — neutral, happy, excited, calm, serious, casual, confident
- Pacing modes — natural, conversational, presentation, tutorial, narrative
- Playback speed and volume set per call
- WAV, MP3, FLAC and OGG output
- Streaming synthesis and playback, and batch processing across multiple text segments
- A queue with real-time status, so concurrent requests can be tracked
- Audio files saved, listed and organised, with two resources covering voice capabilities and usage examples
Requirements
Node.js 18 or newer with a build step for the local route, or a Smithery deployment if you would rather not run it yourself. No model downloads either way.
Setup effort
Build from source — clone the repository and build it, then point your client at the binary