Custom Voices
Voice settings
VoiceConfigure the text-to-speech voice Myra uses on calls. Choose a provider, pick a voice, set the language, and adjust speaking speed. Some providers offer additional tuning controls.
Voice providers
Supported provider catalogs
Azure
156 voices
ElevenLabs
30 voices
Cartesia
21 voices
OpenAI
6 voices
Vomyra
6 voices
Mistral
2 voices
Settings
Core voice configuration fields
| voice_provider | The TTS provider to use for Myra. azureelevenlabscartesiaopenaivomyramistral |
| voice_name | The specific voice to use from the selected provider. Each provider uses its own identifier format. Azure uses a locale plus name, such as en-US-EmmaNeural. ElevenLabs uses a voice ID string. Cartesia uses a UUID. OpenAI and Vomyra use short names like nova or onyx. |
| voice_language | The language code for the TTS voice output. Cartesia and ElevenLabs use short codes like en or hi. Azure uses locale codes such as hi-IN or en-US. |
| voice_speed | Speech speed multiplier. 1.0 is normal pace. Values below 1.0 slow the voice down, and values above 1.0 speed it up. Cartesia supports a range of 0.25 to 4.0. Check your provider documentation for supported ranges. |
Provider-specific tuning
Only supported fields affect the selected provider
Provider-specific fields are ignored by other providers. For example, ElevenLabs stability does not affect Cartesia, Azure, OpenAI, Vomyra, or Mistral voices.
stability
Controls voice consistency. Range: 0.0 to 1.0. Higher values produce more stable delivery. Lower values introduce more expressive variation. Default 0.75.
similarity_boost
Controls how closely output matches the original voice sample. Range: 0.0 to 1.0. Higher values stay closer to the source voice. Default 0.8.
emotion
Sets the emotional tone of the voice output. Supported values include neutral, happy, sad, and angry.
Note