Introducing Eleven Realtime: a simpler way to build natural voice agents

ElevenLabs has introduced Eleven Realtime, speech-to-speech models that listen and answer directly in voice, making it simpler to build natural voice agents. It comes in two versions, Realtime v1 Mini and Realtime v1 Max, with more than 90 languages and over 17,000 voices. The API is compatible with OpenAI's Realtime API to ease migration.

See original publication↗ElevenLabs (blog officiel) · elevenlabs.io