Soniox TTS is a real-time streaming text-to-speech provider on the default Edesy runtime. Each of its voices speaks 60+ languages, and it handles text that mixes English with an Indian language. Use it with Soniox STT if you want one vendor for listening and speaking.
At a glance#
| Item | Value |
|---|---|
| Provider ID | soniox (alias soniox_tts) |
| Runtime | Default runtime |
| Voices | Maya (recommended default), Nina, Emma, Claire, Grace, Mina, Lucia, Daniel, Noah, Jack, Adrian, Owen, Kenji |
| Languages | 60+ per voice |
| Code-switching | Handles mixed English and Indian-language text |
| Your own key (BYOK) | Supported with a Soniox key |
Set it up#
- Open the agent and go to Voice → Text-to-speech.
- Choose Soniox.
- Pick a voice. Maya is a warm default.
- Save and run a test call.
{
"ttsProvider": "soniox",
"ttsVoice": "Maya"
}
FAQs#
Do Soniox voices change per language? No. The same voice speaks every supported language; the model adapts to the text.
Is there an add-on charge? Soniox TTS is a premium voice. See provider costs.
Can one Soniox key cover STT and TTS? Yes. Add your Soniox key once under BYOK.