Telnyx AI Assistants can now appear on video. Six preset avatars, rendered by a model that runs on Telnyx infrastructure, are available in beta.
avatar=true. Saving an avatar doesn't turn every conversation into a video session.Adding a face to a voice agent used to mean adding a vendor: a third-party avatar service sitting between your assistant and the caller, with its own API, its own latency, and its own contract. The avatar model runs on Telnyx infrastructure instead, next to your assistant's voice, transcription, and tools. Video becomes a setting on the assistant you already have, not another integration to maintain. One assistant, one control plane, audio and video on the same network.
avatar_settings when creating or updating an assistant through the API:{ "avatar_settings": { "provider": "telnyx", "model": "foxtrot-1", "avatar_id": "elena", "background_id": "studio" } }
avatar=true; video is delivered over WebRTC. Complete the avatar setup within 60 seconds of connecting.Learn more in the Avatars docs.