openai cs agents demo

Report Abuse

Basic Information

This repository is a demonstration application that showcases a customer service agent interface built on top of the OpenAI Agents SDK. It combines a Python backend that implements agent orchestration logic and example agents with a Next.js frontend that visualizes the orchestration process and provides an interactive chat interface. The demo implements a set of specialist agents and a triage agent that routes user requests to appropriate handlers and enforces guardrails to keep conversations on-topic. It includes example airline-focused flows such as seat changes, flight status inquiries, cancellations, and FAQ handling to illustrate routing and tool usage. The project is intended as a learning and experimentation resource for developers who want to see how to structure, run, and customize multi-agent customer service workflows using the Agents SDK.

Links

App Details

Features
A Python backend that implements the Agents SDK customer service example and orchestrates multiple specialist agents. A Next.js user interface that visualizes agent orchestration steps and provides a chat UI for interactive testing. Example agent types including a triage agent, seat booking agent, flight status agent, cancellation agent, and FAQ agent. Demonstrated demo flows showing routing between agents and typical airline customer service scenarios. Guardrail enforcement with visual indicators for relevance and jailbreak attempts. Simple setup and run instructions for backend and frontend, environment variable configuration for the OpenAI API key, and a modular structure designed to be extended and customized. Licensed under MIT for reuse.
Use Cases
This demo helps developers and teams understand how to compose and orchestrate multiple agents for customer service use cases using the OpenAI Agents SDK. It provides concrete examples of intent triage, specialist agent routing, responses for common airline scenarios, and guardrails that prevent off-topic or instruction-leaking responses. The visual frontend makes it easier to observe the orchestration process and guardrail triggers during conversations, while the backend code shows how to implement prompts, tools, and routing logic. Clear setup steps and runnable demo flows enable quick experimentation and customization, making the repo a practical starting point for building or adapting multi-agent customer support systems.

Please fill the required fields*