DebateMind AI - Multi-Agent Debate System

DebateMind AI is a Streamlit-based Python project that simulates a structured debate between multiple AI agents:

• 🟢 Pro Agent argues in support of the topic.
• 🔴 Anti Agent argues against the topic.
• 🧠 Judge Agent evaluates both sides and delivers a balanced conclusion.

The project uses the Groq API through a clean modular architecture.

Project Structure

debatemind/
├── app.py
├── agents/
│   ├── anti_agent.py
│   ├── judge_agent.py
│   └── pro_agent.py
├── core/
│   ├── debate_engine.py
│   └── llm.py
├── utils/
│   └── prompts.py
├── requirements.txt
└── README.md

Requirements

• Python 3.9+
• Groq API key

Install Python packages:

pip install -r requirements.txt

Groq Setup

1. Create a .env file in debatemind/ (or copy .env.example):

GROQ_API_KEY=your_groq_api_key_here
GROQ_MODEL=llama-3.3-70b-versatile
GROQ_FALLBACK_MODELS=llama-3.1-8b-instant

2. Or export these variables in your shell instead of using .env.

Run The App

Ensure you are in the project root (debatemind/):

streamlit run app.py

How It Works

1. User enters a debate topic (for example, Is AI dangerous?).
2. Debate engine runs for 2-3 rounds.
3. In each round:
   • Pro Agent makes/rebuts an argument.
   • Anti Agent responds/rebuts with context awareness.
4. Judge Agent receives full history and produces:
   • Pro summary
   • Anti summary
   • Strongest points from both sides
   • Balanced final conclusion

Error Handling

The app handles:

• Empty topic input
• Groq connection/auth failures
• Timeout/invalid API responses
• Unexpected runtime errors without crashing the UI

Notes

• The app is intentionally modular for easy extension (more agents, custom prompts, scoring logic, etc.).
• Debate rounds are constrained to 2-3 for stability and UI clarity.