Choose the voice and transcription settings

Voice settings control speech output, while speech-to-text settings control how the caller is transcribed. The language model decides the response. These are separate choices.

Navigation pathAgents → select an agent → Voice

PrerequisitesA voice service available to the selected provider.

Usage steps

Choose an available voice source. For a Native Agent, the editor offers ElevenLabs and Native audio.
Select the connected integration where required. Choose a voice and a compatible TTS model, then review the voice Advanced options if needed.
In Speech-to-Text, choose the STT Provider and model, then set the primary Language.
Add brand names, product names or specialist terms to Key Terms. Press Enter or a comma to add a term.
Choose the language-model provider and model. Save with Update Agent and test pronunciation, turn-taking and recognition.

Choose the voice and transcription settings — Figure 8. Voice, TTS model, transcription and language-model settings are configured in the Voice tab.

Expected result. The agent speaks with the selected voice and transcribes the intended language.

Practical notes and troubleshooting. If the current voice is not found in the connected key, check the integration or choose another available voice. Provider and model choices vary.

Browse and preview voicesVoice Library helps you compare catalogue voices before choosing one for an agent. Filter the catalogue, listen to samples, and keep a shortlist in Favourites.Test with a browser callWeb Call lets you speak to an agent from your browser. Use it to hear the opening message, test responses and check the behaviour of configured tools.