Voices

Overview

Soniox Text-to-Speech offers a curated set of studio-quality voices, each designed to sound natural, expressive, and consistent across languages.

• Any voice, any language → Every voice works with all 60+ supported languages, pick a voice once and keep the same speaker across your whole product.
• Consistent identity → The same voice preserves its timbre and style whether it is speaking English, Japanese, or Spanish.

Voices vary in gender, age, energy, and accent, so you can match the speaker to your product's personality.

Usage

Set the voice by passing its name in the voice field of your Text-to-Speech request:

{
  "model": "tts-rt-v2",
  "language": "en",
  "voice": "Daniel"
}

Swap language to any supported ISO code to have the same voice speak a different language, the speaker identity stays consistent.

See models for the currently available Text-to-Speech models.

Listing available voices

To retrieve the voices available for each model, use the Get TTS models endpoint. Each model in the response includes its voices, with the name to pass in the voice field, the gender, and a short description of the voice's character:

{
  "id": "Mina",
  "description": "A steady, clear voice with a natural presence and measured delivery that feels confident, warm, and easy to listen to.",
  "gender": "female"
}

Custom voices

Need a voice that isn't built in? With voice cloning you can create your own voice from a short reference clip and use it by its voice ID in the voice field, just like a built-in voice.