MetaVoice
A full-duplex speech AI model designed for natural, human-like revenue and sales calls.
Category: AI Voice

Overview
MetaVoice is a speech-to-speech AI model that enables voice agents to handle interruptions, overlaps, and background noise in real-time. Unlike traditional cascaded systems, it processes conversation as a single model to ensure natural, fluid interaction during revenue-generating calls.
How to use it?
Users define the agent’s workflow, personality, and tool integrations via prompts. The model can be deployed within a private cloud (VPC) to ensure data security and can be coached or fine-tuned using call feedback.
Features
Full-duplex conversation, Real-time interruption handling, On-premise deployment, Customizable personality, Built-in control guardrails, Detailed call observability