tasty-wheel is an automated options trading engine designed to execute the Cash-Secured Put (CSP) "wheel" strategy via the Alpaca API with strict safety mechanisms. Built around deterministic execution, the system decouples trade logic from generative AI: trading decisions and risk assessments are governed purely by hardcoded mathematical gates and predefined criteria to eliminate hallucination risks and unexpected portfolio behavior. To bridge the gap between automated execution and transparency, the architecture integrates a read-only Large Language Model (LLM) narrator. This narrator observes market data, executed trades, and rule checks in real time, generating clear, human-readable explanations of why specific trades were opened, held, or skipped without ever being in the active trade path. By pairing rigid algorithmic discipline with AI-driven explainability, tasty-wheel provides traders with an automated, auditable, and risk-controlled environment for systematically executing options income strategies.
Category tags: