What do you need help with?
Feature request
Issue area
Custom Voice / ElevenLabs
Impact
Blocker - I cannot continue
Character ID
f059089a-6272-11f1-880b-42010a7be02e
Where does it happen?
API / Backend
Voice provider
Elevenlabs
Voice name
whatever
Short summary
Custom voice using new elevenlabs api
What happened?
We use an ElevenLabs private cloned voice through the Convai ElevenLabs API integration.
ElevenLabs released eleven_v4_turbo on September 28, 2026. It is specifically designed for realtime agents and provides approximately 100 ms median inference latency.
We would like our Convai character to use:
model_id = eleven_v4_turbo
I understand that Eleven v4 Turbo uses the new ElevenLabs Text-to-Dialogue WebSocket:
wss://api.elevenlabs.io/v1/text-to-dialogue/stream-input
rather than the legacy Text-to-Speech WebSocket used by Flash/Multilingual models.
Does Convai currently support this endpoint/model for private ElevenLabs voices?
If not, could this be enabled for our account or added to the ElevenLabs integration?
I checked the relevant documentation, resources, and similar forum posts before posting.
on
I confirm that I did not include API keys, tokens, passwords, secrets, private character data, private conversation logs, or customer data.
on