AI voice agent
An AI voice agent is an AI agent that talks to people by voice only, over the phone or in an app. It hears the person, decides what to say with a language model and answers in a synthetic voice, with no visual face.
How AI voice agents work
A voice agent chains speech recognition, a language model and text-to-speech, with turn detection to know when the person has finished speaking. Platforms such as LiveKit carry the audio in real time.
Adding a real-time avatar to the same pipeline turns a voice agent into a video agent, without changing its logic.
How Beyond Presence does it
The Beyond Presence Speech-to-Video API adds a face to the voice agent you already run: you keep your speech recognition, LLM, voice and orchestration, and the Avatar renders the audio as video in about 100 ms. LiveKit is the recommended integration. Bunjee measured 2.5x longer engagement with an Avatar than with its voice-only agent. Read the AI agent stack in 2026.
Frequently asked questions
How does Beyond Presence turn a voice agent into a video agent?
The Speech-to-Video API takes your agent's audio and returns a synchronised Avatar in about 100 ms, so people see a face as well as hear a voice.
Do I need to change my voice agent's stack?
No. Your LLM, voice and orchestration stay as they are. LiveKit is the recommended integration.
What difference does a face make?
Beyond Presence customer Bunjee measured about 2.5x more engagement per conversation with an Avatar than with a voice-only agent. That is one customer, not a controlled test.