Cartesia

Cartesia provides text-to-speech for Vapi voice agents.

You can use Cartesia through Vapi’s default integration, or connect your own account in Integrations.

Supported capability

Text-to-speech

Set voice.provider to cartesia and voice.voiceId to a voice. voice.model is optional because Vapi selects the appropriate model for the voice ID when you omit it. To select a model explicitly, use one of the supported model IDs below.

Cartesia offers many voices and supports voice cloning. You can browse and copy a voice’s ID from the Voice Library.

$
curl -X PATCH "https://api.vapi.ai/assistant/ASSISTANT_ID" \
>
  -H "Authorization: Bearer $VAPI_API_KEY" \
>
  -H "Content-Type: application/json" \
>
  -d '{
>
    "voice": {
>
      "provider": "cartesia",
>
      "model": "sonic-3.5",
>
      "voiceId": "VOICE_ID"
>
    }
>
  }'

For additional API configuration options, review the CartesiaVoice fields in the Create Assistant API reference.

Supported models

Supported languages

Language support depends on the selected model. The following model and language combinations are currently supported by Vapi. Dated snapshot models (for example, sonic-3-2026-01-12) match their base model’s language coverage.

Related