# GPT Realtime Voice App: I Replaced Customer Service With 2 AI Agents

Source: https://www.youtube.com/watch?v=b22HaCm5dHY
Recap page: https://rapidrecap.app/video/b22HaCm5dHY
Generated: 2025-09-02T00:02:05.741+00:00

---
## Quick Overview

OpenAI introduces new gpt-realtime and Realtime API updates for production voice agents, demonstrating enhanced reasoning and more natural speech, with a two-agent system (Chat Agent and Supervisor Agent) for complex tasks and a pricing reduction of 20% for gpt-realtime.

**Key Points:**
- OpenAI releases gpt-realtime and Realtime API updates for production voice agents, offering improved reasoning and natural speech capabilities.
- The system utilizes a two-agent architecture: a 'Chat Agent' (gpt-4o-realtime-mini) for basic interactions and a 'Supervisor Agent' (gpt-4.1) for complex tasks requiring tool calls.
- The gpt-realtime model achieves 66.5% accuracy in ComplexFuncBench Audio Function Calling and 30.5% in MultiChallenge (Audio) Instruction Following, outperforming previous models.
- Users can run the demo locally by cloning the OpenAI Realtime Agents repository, installing dependencies, setting their OpenAI API key, and running 'npm run dev'.
- The Realtime API and gpt-realtime model are available to all developers, with a 20% price reduction compared to the gpt-4o-realtime-preview model.
- The OpenAI Agents SDK provides a unified interface for defining agent behaviors, tool integrations, orchestration, state management, and low-latency streaming interactions.

![Screenshot at 00:01: The video opens with a title card announcing 'Introducing gpt-realtime and Realtime API updates for production voice agents', setting the stage for the demonstration.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-00-01.png)

**Context:** This video introduces OpenAI's advancements in voice agent technology with the release of gpt-realtime and Realtime API updates. It showcases a demonstration of these new capabilities, highlighting improved natural language processing and reasoning for voice agents. The video explains the underlying architecture, which uses a combination of a 'Chat Agent' for simpler tasks and a 'Supervisor Agent' for more complex operations involving external tools, and provides a practical guide on how developers can set up and run the demo locally.

## Detailed Analysis

OpenAI has released significant updates to its Realtime API and introduced the gpt-realtime model, enhancing the capabilities of production voice agents. The core improvement lies in the model's stronger reasoning and more natural speech generation, allowing it to handle complex interactions more effectively than previous models. The system employs a dual-agent approach: a 'Chat Agent' (gpt-4o-realtime-mini) manages basic conversational flow, while a 'Supervisor Agent' (gpt-4.1) takes over for more complex tasks, including the ability to call external tools. This architecture is visualized in a sequence diagram, showing how user messages are processed and routed between agents and tools. Performance benchmarks are presented, demonstrating that gpt-realtime achieves 66.5% accuracy in ComplexFuncBench Audio Function Calling and 30.5% in MultiChallenge (Audio) Instruction Following, surpassing older models. The video also provides a step-by-step tutorial for developers to run the demo locally, involving cloning the OpenAI Realtime Agents repository from GitHub, installing Node.js packages using 'npm install', setting the OpenAI API key as an environment variable, and running the development server with 'npm run dev'. The pricing for gpt-realtime has been reduced by 20% compared to the preview version, making it more accessible. The OpenAI Agents SDK is highlighted as a toolkit for building, managing, and deploying advanced AI agents, offering features like a unified interface, built-in orchestration, state management, and easy integration with the Realtime API for low-latency streaming.

### Introduction

- OpenAI announces gpt-realtime and Realtime API updates
- Enhanced reasoning and natural speech
- Two-agent system (Chat Agent and Supervisor Agent)
- Pricing reduction for gpt-realtime

### Agent Architecture

- Chat Agent (gpt-4o-realtime-mini) for basic tasks
- Supervisor Agent (gpt-4.1) for complex tasks and tool calls
- Sequence diagram illustrating agent interaction and tool usage

### Performance Benchmarks

- gpt-realtime achieves 66.5% accuracy in ComplexFuncBench Audio Function Calling
- 30.5% accuracy in MultiChallenge (Audio) Instruction Following
- Outperforms previous models

### Local Setup Tutorial

- Clone repository from GitHub
- Install dependencies with 'npm install'
- Set OpenAI API key
- Run development server with 'npm run dev'

### Pricing and SDK

- 20% price reduction for gpt-realtime
- OpenAI Agents SDK provides tools for building, managing, and deploying agents
- Features include unified interface, orchestration, state management, low-latency streaming

### Output Guardrails

- Assistant messages checked for safety and compliance
- Guardrail logic in src/app/App.tsx
- Handling of message states (IN_PROGRESS, FAIL, PASS)
- Customization of moderation triggers

![Screenshot at 00:00: The video begins with an introduction to OpenAI's new gpt-realtime and Realtime API updates for voice agents.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-00-00.png)
![Screenshot at 00:05: A demo interface shows a conversation with a voice agent, displaying transcript and logs.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-00-05.png)
![Screenshot at 00:25: A sequence diagram illustrates the interaction between User, Chat Agent, Supervisor Agent, and Tool.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-00-25.png)
![Screenshot at 00:30: The video highlights the 'About the OpenAI Agents SDK' section, detailing its features.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-00-30.png)
![Screenshot at 00:36: A function call result shows retrieved user account information, including name, phone, and billing details.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-00-36.png)
![Screenshot at 01:07: A bar chart compares the accuracy of 'gpt-realtime' with previous models in audio function calling.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-01-07.png)
![Screenshot at 01:10: Another bar chart shows the accuracy of 'gpt-realtime' compared to previous models in audio instruction following.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-01-10.png)
![Screenshot at 01:14: A map of Seattle is displayed, likely demonstrating a real-world application of the voice agent.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-01-14.png)
![Screenshot at 01:35: A terminal window shows the command 'git clone' to download the OpenAI Realtime Agents repository.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-01-35.png)
![Screenshot at 01:43: The terminal displays 'npm i' to install project dependencies, followed by audit information about packages and vulnerabilities.](https://ss.rapidrecap.app/screens/b22HaCm5dHY/00-01-43.png)