Multi-Agent Debate

AI Debate Simulation Platform

Multi-Agent Debate is a platform that simulates structured debates between AI agents arguing opposing sides of any topic. A user submits a topic, which the system normalizes into a formal resolution and enriches with web-based evidence. Two AI agents then take opposing positions and work through a complete debate cycle: opening statements, crossfire, rebuttals, and closing arguments, with results streaming live to the user interface.

The backend is built with FastAPI and SQLite, orchestrating the debate flow and managing agent state across rounds. The primary AI model is Groq API, with Google Gemini as a fallback provider to ensure reliability. A web search integration supplies agents with real evidence to ground their arguments, and a text-to-speech feature brings the debate to life audibly.

The frontend is a React + TypeScript app built with Vite, consuming the streaming API responses to render each debate segment as it is generated. The application is deployed on Vercel, making the full debate experience accessible online.



WebApp GitHub Repo