In simple English
This project shows how an airline support desk can run on automated helpers. A main helper routes messages to specialists for flights, seat changes, cancellations, refunds, and policy questions. You can watch the system switch helpers and reply in real time. It requires an API key from OpenAI, which charges fees based on how much you use it. Setting it up on your computer takes a few minutes of terminal work.
What it is
A full-stack demonstration application containing a Python backend for agent orchestration and a Next.js user interface. It illustrates multi-agent workflows, tool calls, and built-in safety guardrails.
What it is good for
Solo builders who want to learn how to build multi-agent customer service applications using official OpenAI tools.
- Includes a modern chat interface built with Next.js and ChatKit.
- Demonstrates complex multi-agent handoffs and specialized tools.
- Shows how to implement guardrails for relevance and safety.
- Requires an active OpenAI API key, which incurs usage costs.
- Needs two separate terminal processes running simultaneously for backend and frontend.
- Designed strictly as a demonstration rather than a production-ready helpdesk.
Writing a custom multi-agent customer service orchestration loop from scratch.
Next.js
- openai/openai-agents-python30k stars
The underlying Python library used by this demo for agent orchestration.
- deepset-ai/haystack27k stars
A Python framework for joining language models, document search and tools into one working pipeline.
- Giskard-AI/giskard-oss5.9k stars
A Python library that writes tests for language-model programs and tries to break them on purpose.
- ggml-org/llama.cpp132k stars
A C and C++ library and command-line tool for running language models locally.
cd python-backend python -m venv .venv source .venv/bin/activate pip install -r requirements.txt cd ../ui npm install