MadMoon

MadEye is an AI-powered geopolitical threat intelligence system that monitors 62 data collectors across 5 languages (Arabic, Farsi, Russian, Chinese, English) to predict attacks on critical energy infrastructure before they happen. The system detected the October 2026 Strait of Hormuz tanker mine incidents 12 hours before markets priced them in, using real-time OSINT from Houthi military Telegram channels, Iranian state media, satellite thermal data, and shipping traffic analysis. MadEye runs on a single AWS instance and an 8GB MacBook Air. No GPU cluster. No enterprise infrastructure. Built by one developer. The system uses Llama for multi-language analysis of raw Arabic and Farsi signals, classifying threat levels from unstructured text across 12 specialized scrapers. Each scraper produces a confidence-weighted score that feeds into a unified threat assessment. 62 collectors. 12 scrapers. 5 languages. Real predictions. One person.