Text to Speech

Voices and languages

Listen to the available voices and select one for your request.
View as Markdown

Choose a voice with the voice field. If omitted, english_female_professional is used.

Available voices

Both samples were generated with Qwen3-TTS Base using the same text: “Hello, welcome to Nari Labs.”

English Female Professional

english_female_professional · English · Default

English Female Articulate

english_female_articulate · English

Select a voice

Pass the ID in a speech generation request:

1{
2  "model": "qwen3-tts:free",
3  "input": "Your order has shipped.",
4  "voice": "english_female_articulate",
5  "language": "en"
6}

An unknown voice ID is rejected. Display names are labels; use the exact ID in requests.

Language selection

Both current voices speak English.

languageBehavior
enSelect English explicitly
auto or omittedUse the selected voice’s assigned language: English

auto does not detect or translate the input. Other language values are rejected. English numeric expressions are normalized before synthesis.