🤖 Multi-Agent Chainlit Assistant

Welcome to the Multi-Agent Chainlit Assistant — a smart assistant framework built using the OpenAI Agents SDK and Chainlit UI, with agent routing powered by a decision-making agent.

🚀 Built with love using uv, Chainlit, and OpenAI Agents SDK.

---

📦 Project Overview

This assistant uses a main agent to interpret and route user queries to one of the following tools:

• 🧠 Main Agent — smart router that plans and decides.
• 😈 Shaitani Calculator Agent — handles all math problems but always gives confidently wrong answers for fun.
• 🌐 Web Search Agent — uses real-time tools to answer factual and current questions.

The interface is built with Chainlit to provide a live interactive UI.

---

🧪 Features

• 🔁 Multi-agent coordination (math + search)
• 🧮 Fun, incorrect calculator agent
• 🌐 Real-time web information retrieval
• 🧠 Autonomous planning via main agent
• 💡 Prompt engineering built-in for realistic agent behaviors

---

🧰 Tech Stack

Tool	Purpose
uv	Fast Python package manager
OpenAI Agents SDK	Build intelligent LLM agents
Chainlit	UI framework for LLM assistants
dotenv	Environment variable handling
Python 3.11+	Language

---

🚀 Quick Start

1. Clone the repository

git clone https://github.com/ZohaibCodez/multi-agent-chainlit-assistant.git
cd multi-agent-chainlit-assistant

2. Initialize & Activate Environment (with uv)

uv init  # initializes pyproject.toml
uv venv  # creates a virtual environment
source .venv/bin/activate  # activate (Linux/Mac)
# OR
.venv\Scripts\activate  # activate (Windows)

3. Install Dependencies

uv add openai-agents chainlit dotenv

4. Set Environment Variables

Create a .env file:

GEMINI_API_KEY=your_gemini_key_here

---

5. Run the App

uv run chainlit run main.py -w

• App will be available at: http://localhost:8000

---

📁 Project Structure

multi-agent-chainlit-assistant/
│
├── agent_files/             # Agent implementations
│   ├── main_agent.py       # Main routing agent
│   ├── shaitani_agent.py   # Fun calculator agent
│   └── web_search_agent.py # Web search capabilities
│
├── config/                  # Configuration files
│   ├── agents_config.py    # Agent settings
│   └── prompt_templates.py # Agent prompts
│
├── tools/                   # Custom tools
│   ├── shaitani_calculator_tool.py
│   └── web_search_tool.py
│
├── chainlit.md            # Welcome screen content
├── main.py               # Application entry point
├── pyproject.toml        # Project metadata & dependencies
└── README.md            # This file

---

🧠 Agents Overview

🧠 Main Agent

• Routes queries to appropriate tools
• Plans and reflects on outcomes
• Handles multi-part tasks

😈 Shaitani Calculator Agent

• Intentionally gives wrong answers to all math queries
• Overconfident and funny personality

🌐 Web Search Agent

• Uses real-time tools to answer factual queries

---

📌 Notes

• Designed for fun and experimentation with agent routing logic
• Based on OpenAI's new multi-agent architecture
• Chainlit enables smooth visual interaction

---

🛠 Future Improvements

• Add memory and context tracking
• Include user authentication and logging
• Add support for more tools like PDF reader, CSV interpreter, etc.

---

👤 Author

Developed By Zohaib Khan

---

📄 License

MIT License © 2025 Zohaib Khan.