Skip to main content
Welcome to the Speech-to-Speech (S2S) Agent Creation API, powered by our model ROSE.
If you’re building human-like voice AI agents that can listen, think, and respond naturally, you’re in the right place.

​
Welcome

Meet ROSE, our real-time Speech-to-Speech model built for next-generation voice AI agents.
Unlike traditional text-based systems, ROSE allows direct voice-to-voice interaction, enabling agents that sound expressive, react instantly, and feel more human. This documentation will guide you through every step of building your own voice AI agent, from creation to live interaction.

​
Getting Started

Building your voice agent with ROSE is simple:
  1. Authenticate — Secure your API calls with a bearer token.
  2. Create an Agent — Use our /agents endpoints to create and manage your voice AI agents.
  3. Connect via WebSocket — Communicate with ROSE in real time using WebSocket connections.
    • You can connect over web or telephony to exchange live audio streams.

​
Authentication

Before making API calls, you’ll need an API key.
You can generate one by logging in to your AIvoco Dashboard. 
Authorization: Bearer <YOUR_API_KEY>