Smart attendance systems have become increasingly important for educational institutions and workplaces, where accurate and efficient attendance management is essential. Conventional attendance methods, including manual registers and RFID-based systems, are prone to human error, proxy attendance, and increased infrastructure costs. This paper presents a standalone Edge AI-based smart attendance system using the ESP32-CAM module for real-time facial recognition in Internet of Things (IoT) environments. The proposed system integrates image acquisition, face detection, and recognition directly on an embedded platform, eliminating the need for external computing resources or continuous internet connectivity. A compact custom-designed printed circuit board (PCB) incorporating power management circuitry, a rechargeable battery, and a charging module enables portable, low-power, and reliable operation. The embedded vision algorithm captures facial images, performs on-device recognition, and records attendance in real time, thereby reducing processing latency and enhancing system efficiency. Experimental evaluation demonstrates reliable face detection and recognition performance in controlled indoor environments while maintaining low power consumption and cost-effectiveness. The proposed solution offers a scalable and practical attendance management system for educational institutions, offices, and other resource-constrained environments. Future work will focus on improving recognition accuracy under varying illumination and pose conditions, integrating liveness detection for enhanced security, and enabling secure cloud synchronization to support large-scale deployments.