An intelligent system that orchestrates a debate between two AI models to refine answers and reach consensus. Perfect for tackling math, logic, fact-checking, or coding challenges with enhanced reliability through collaborative reasoning.

🗣️ DebatingAgents

An intelligent system that orchestrates a debate between two AI models to refine answers and reach consensus. Perfect for tackling math, logic, fact-checking, or coding challenges with enhanced reliability through collaborative reasoning.

🖼️ Snapshot

🔧 Prerequisites

• Python 3.13 or higher
• uv (efficient Python package management)
• OpenAI API key
• Google Claude API key

📥 Installation

• Clone the project

git clone https://github.com/smaranjitghose/DebatingAgents

• Navigate inside the directory

cd DebatingAgents

• Create a .env file and store the API keys from OpenAI and Google Gemini as:

OPENAI_API_KEY=your-openai-api-key-here
GEMINI_API_KEY=your-gemini-api-key-here

• Run the agent using uv (on the first run, it will create a virtual environment and install dependencies)

uv run main.py

• Start the debate with examples:

🗣️ You: How many prime numbers are less than 100?

💡 Usage Examples

• Math Problems How many integers between 1 and 1000 are divisible by 3, 5, or 7 but not 3 and 5?

• Logic Puzzles You found a treaure chest at the bottom of the Indian Ocean. There are 60 diamonds, twice as many rubies and one-third as many emeralds as rubies. What are the number of diamonds, rubies and emeralds you have?

• Programming Assistance How many lines of code are needed for a basic binary search algorithm?

🔍 How It Works

• API Integration 🌐 - Connects to OpenAI and Gemini APIs using secure environment variables.
• Response Validation ✅ - Uses Pydantic to ensure structured responses (integer answer and explanation).
• Debate Protocol ⚔️