Text-to-Speech
Text-to-Speech
Generate natural speech in 60+ languages over a WebSocket or REST, with voice cloning and emotion control.
Soniox Text-to-Speech turns text into natural speech in 60+ languages. Stream text in over a WebSocket and start playing audio after the first few words, or generate a whole file with one REST request. Names, numbers, and email addresses come out the way people say them.
Get started
Generate your first audio in a few minutes.
Real-time API
Stream text in and audio out with low latency.
REST API
Generate a full audio file in one request.
Voices
Browse the built-in voices and pick one.
Voice cloning
Create a voice from a short recording.
Emotion & tone
Make speech sound warm, serious, or cheerful.