Agora Quantum

Agora Quantum is an experiment in fully autonomous investing: what happens when you let AI agents manage a portfolio with zero human intervention? The answer is a 9-agent debate system that scans 8,000+ securities daily.

Multi-Agent Architecture

9 specialized agents with distinct personalities and investment philosophies:

• Value investors looking for underpriced assets
• Momentum traders riding trends
• Contrarians betting against the crowd
• Risk managers ensuring portfolio stability
• Macro analysts watching economic indicators

The Debate Protocol

3 rounds of structured debate where agents argue for and against positions. Each agent must defend their thesis against challenges from others. Final decisions require consensus across multiple investment philosophies.

Portfolio Construction

Black-Litterman optimization combines agent views with market equilibrium. This produces portfolios that reflect AI conviction while maintaining risk-adjusted returns. Position sizing adapts to confidence levels.

Autonomous Execution

Trades execute automatically via Alpaca API. No human approval required. The system handles market orders, limit orders, and position management 24/7.

Infrastructure

Runs entirely on a Raspberry Pi, proving that sophisticated AI trading doesn't require expensive infrastructure. PostgreSQL for state, Prometheus for monitoring, Docker for deployment.

The Experiment

This isn't about beating the market (though that would be nice). It's about exploring what autonomous AI systems can do when given real-world agency with real consequences.