This document describes an IoT-based facial recognition system for access control. The system uses a Raspberry Pi connected to a camera to capture facial images and recognize people to control access to homes and offices. It works by training a face recognition model on a dataset of images, then using that model to identify people in new images and open doors or grant access. The system aims to provide secure and hands-free access while reducing costs compared to traditional locks. Some limitations are accuracy depends on lighting and image quality, but future work could improve the system by adding a second camera for 3D analysis and increasing the recognition threshold.