Voice Agent Builder

The Voice Agent Builder is the environment for creating phone-based AI agents that handle inbound and outbound calls.

Voice Agent Builder — overview
A Voice Agent uses speech-to-text (STT) to transcribe what a caller says, passes the transcript to a language model (LLM) for reasoning, and speaks the response back using text-to-speech (TTS) — all in real time over a telephony connection.

Use Voice Agent when you need:

  • Phone-based interactions — the agent must receive or place calls via a phone number.

  • Real-time speech processing — conversations happen over voice, not text.

  • Outbound campaigns — the agent places calls to a list of contacts on a defined schedule.

  • Inbound call handling — the agent answers incoming calls and routes or resolves them autonomously.

For text-based chat agents, use the Chatbot Builder instead. For multi-step reasoning workflows over APIs and data sources, use the Agentic Flow Builder.