Security in public spaces such as internet cafés (warnet) is a major concern, especially during non-operational hours. This study develops a real-time security system based on the Internet of Things (IoT) by integrating the ESP32 and ESP32-CAM microcontrollers to enhance surveillance effectiveness in warnet environments. The system uses a PIR sensor to detect motion, then automatically sends a notification to the Telegram application and triggers image capture via the ESP32-CAM. In addition to automatic detection, users can also manually control the system via Telegram, such as activating the buzzer, requesting snapshots, or accessing real-time video streaming using ngrok. The system was developed using the Waterfall method, which includes requirement analysis, design, implementation, and testing. The results show that all system features operate properly, are responsive, and support efficient remote monitoring. With its low cost and optimal technology integration, the system is a viable alternative to conventional security systems for small-scale businesses.