Multi-Agent AI Support System

This project focuses on designing and incrementally building a modern AI-powered customer support system using a scalable full-stack architecture. The aim is to demonstrate how a real-world AI system can be constructed step by step—starting with core logic and UI, and gradually integrating advanced technologies.

---

Project Goal

To build a multi-agent AI system that intelligently routes user queries to specialized agents (Support, Billing, Order, etc.) using a modern, production-grade tech stack.

---

Current Implementation

Frontend

• HTML
• CSS
• Vanilla JavaScript
• Real-time chat UI
• Typing/streaming-style responses

Backend

• Node.js
• Hono.dev
• Modular multi-agent architecture
• Intelligent routing logic
• RESTful APIs

---

Features Implemented

• Multi-agent routing
• Support, Billing, Order agents
• Chat-based UI
• Streaming-style typing effect
• API-based communication
• Modular backend structure

---

Architecture Overview

1. User sends a message from the frontend UI.
2. Message is sent to the backend API.
3. Router Agent analyzes the message.
4. The appropriate specialized agent is selected.
5. The agent generates a response.
6. The response is streamed back to the frontend.

---

API Endpoint

Send Message

POST /api/chat/messages

Request Body

{
  "message": "Your question"
}

Response

{
  "agent": "support",
  "reply": "Generated response",
  "history": []
}

---

Setup Instructions

Backend

npm install
node index.js

Server runs at:

http://localhost:3000

---

Frontend

Open index.html using a browser or Live Server.

---

Author

Built by Juveriya Anjum Shaik

---

License

MIT License