ElevenLabs introduces Eleven Realtime, a speech-to-speech model for voice agents
ElevenLabs announced Eleven Realtime, a speech-to-speech model class for voice agents. The model is available in ElevenAgents and through the Realtime API.
Why it matters
Builders of voice agents can use a speech-to-speech model from ElevenLabs without assembling separate speech recognition, language model, and text-to-speech components. Availability in both ElevenAgents and the Realtime API gives a choice between a managed agent product and direct API integration.
Who should care
Developers building voice agents, customer support automation teams, ecommerce brands adding voice ordering or support, agencies building voice products for clients
What you can do
Test Eleven Realtime through the Realtime API on a sample voice agent flow and measure latency and naturalness against your current speech pipeline.