Choosing Your Agent's Voice

What this guide covers
How to pick your agent's voice, change the speech model that renders it, and fix words the agent mispronounces. All three are on the agent's page in My Agents.

Picking a voice

Open the agent from My Agents.
Find Voice Selection and open the list.
Choose from:
- Default ElevenLabs Voice: the agent’s standard voice.
- Your Cloned Voices: voices you’ve cloned, if any.
- ElevenLabs Premade Voices: a library of ready-made voices, each with a short description.
Select one. It’s saved straight away and used from the next call.

Speech model

Speech Model is the text-to-speech engine that turns the agent’s words into audio. The list shows the models available for the agent’s language. Newer models usually sound more natural; faster ones reduce the pause before the agent replies. If you’re unsure, leave it as it is.

Pronunciations

If the agent says a name or word wrong, add a rule under Pronunciations:

In the first box, type the word exactly as it appears in the agent’s instructions, for example Margie.
In the second box, type how it should sound, for example Marghee.
Click Save. Use the preview to hear it.

Keep the real spelling in your instructions. Pronunciation rules change only how the word is spoken, not what the agent writes in texts or emails.

Video meetings

In TriLuna video meetings the agent runs on a separate system with its own voice, so the voice chosen here applies to phone calls. The speech-recognition language for video is under Video Conference Behavior.

Common questions

The new voice didn’t play on my call. Voice changes apply from the next call. Hang up and call again.

Can I use my own voice? Yes, if you have a cloned voice. It appears under Your Cloned Voices. To have a voice cloned, contact support.