OpenAI Realtime Voice Agents

Production-ready components for voice agent integration

Back to Home
Realtime Voice Session
OpenAI Realtime API Integration
disconnected
Configuration
Model: gpt-4o-realtime-preview
Voice: alloy
VAD: Enabled
Temperature: 0.8
Audio Visualizer
Real-time audio level visualization
Agent Orchestrator
Multi-agent voice workflow management
3 agents
General Assistant
Active

Handles general queries and conversations

searchcalculator
Available Agents
Tool Executor
Function calling during voice conversations
2 tools
Available Tools
searchcalculator
Tool Calls
No tool calls yet
Conversation History
Real-time voice conversation transcript
Hello, can you help me with something?
1:46:15 AM
Of course! I'd be happy to help. What do you need assistance with?
1:46:20 AM1200ms
I need to calculate 25 * 48
1:46:25 AM
Let me calculate that for you. 25 * 48 = 1200
1:46:30 AM800ms

Features

  • •WebRTC and WebSocket transport layers
  • •Real-time audio streaming and visualization
  • •Multi-agent orchestration with handoffs
  • •Function calling with approval workflows
  • •Voice activity detection (VAD)
  • •Conversation history and transcription
  • •Mute controls and audio management
  • •Full TypeScript support with type safety