Supported languages

Overview

Soniox Text-to-Speech AI supports speech generation in 60+ languages with natural, expressive output, all powered by a single, unified AI model.

• Any voice, any language → Every voice can speak in every supported language.
• Consistent quality → The same model delivers uniform pronunciation and prosody across languages.

All languages are available in both:

• Real-time API → Stream generated audio with low latency as text arrives.
• REST API → Synthesize full text inputs in a single request.

To use a language, pass its ISO code from the table below in the language field of your Text-to-Speech request.

To programmatically retrieve the full list of supported languages, use the Get TTS models endpoint.

Note: Some languages support only specific scripts:

• Serbian (sr) and Bosnian (bs): Latin script only (not Cyrillic).
• Kazakh (kk): Cyrillic script only (not Latin).
• Chinese (zh): Simplified Chinese only (not Traditional Chinese).

---

Supported languages