AEGIS v3.1 is an institutional-grade autonomous options trading desk built natively on Alpaca's Trading API. Unlike typical AI trading bots that open directional bets and walk away, AEGIS runs the complete Wheel strategy lifecycle — discover, sell premium, manage, hedge, and close — with zero human intervention. The system orchestrates five specialized agents. The News Agent ingests live Google News RSS headlines to classify market regime and detect earnings blackouts. The Quant Agent screens live option chains, computes Black-Scholes Greeks and IV-rank-adjusted yield, and enforces duplicate-position guards. The Sentiment Agent runs a six-LLM ensemble consensus with automatic failover. The Portfolio Manager runs first every cycle, applying 50% take-profit, 200% stop-loss, ITM rolls, and expiry sweeps while booking realized P&L. The Risk Governance Agent holds sovereign veto power with hard limits and executes real SPY protective puts when its live SPY ATM-IV stress index spikes above 25. Every order, veto, hedge, equity snapshot, and cycle is journaled to SQLite, producing measured analytics: Sharpe ratio, max drawdown, profit factor, and win rate. Operators interact through a real-time WebSocket cockpit, 10 CLI commands, 11 FastMCP tools, Telegram alerts, and AI-generated voice briefings. Team ApexArbitrage: Muhammad Zun U Rain Tahir, Anne Loraine A. Pardillo, Muhammad Abdullah Khalid.
Category tags: