Gas leakages in residential and industrial environments pose a critical threat to life and property due to their flammable and explosive nature. This paper presents the design, development, and evaluation of an IoT-based Gas Leakage Alert System — an integrated technological framework that combines Internet of Things (IoT) sensor networks, edge computing, cloud platforms, and AI-driven visual verification to monitor gas concentration levels, predict threshold breaches, and disseminate life-saving alerts in real time. The system deploys an MQ-6 gas sensor with a NodeMCU (ESP8266) microcontroller to continuously measure LPG/methane concentration in parts-per-million (PPM). Upon exceeding configurable thresholds, the system triggers local alarms, activates a servo-controlled solenoid valve and exhaust fan, and dispatches multi-channel notifications via the Blynk IoT platform, GSM/SMS, and email. An optional YOLO-based visual verification layer reduces false positives through sensor fusion of chemical and image data. Experimental evaluation across 72 hours of continuous operation demonstrated a 95% overall success rate, local detection-to-actuation latency of 0.6 seconds, and cloud notification latency of 3.2 seconds under standard Wi-Fi conditions. Challenges related to sensor calibration drift, humidity-induced reading variability, and rural connectivity gaps are discussed alongside future enhancement pathways including TinyML gas fingerprinting, LoRaMesh networking, and energy harvesting for autonomous deployment.