<![CDATA[The high incidence of home burglaries in Indonesia has driven the need for the development of more modern and effective security systems compared to conventional manual locks, which still have many weaknesses. This research develops an Internet of Things (IoT)-based home door security system by implementing face detection as the primary authentication method. The system is designed using an ESP32 as the main controller, integrated with an ESP32-CAM for face recognition, a 4x4 keypad as a PIN input medium, an SW-420 vibration sensor to detect forced break-in attempts, a buzzer as a warning alarm, and a solenoid door lock as the door-opening and locking actuator. The system operates with two authentication mechanisms, namely face recognition and PIN input, thereby providing layered security. If the authentication process is successful, the door will open automatically and the system will send a notification reading "Door opened, welcome" via the Telegram application to the homeowner. Conversely, if authentication fails or the vibration sensor detects vibrations indicating a break-in attempt, the system will activate the buzzer as a local alarm and send a real-time warning message reading "Danger alert, there is an unauthorized person at the house." Test results show that the system is capable of recognizing registered faces with a good level of accuracy and sending notifications with a delay of less than 3 seconds. With its dual authentication features and real-time notifications, this system is considered capable of significantly improving home security compared to conventional methods.]]>
