Menstrual discomfort significantly affects women’s productivity and engagement in structured environments such as offices, schools, colleges, and conferences. However, current access control systems fail to address this issue, resulting to inadequate accommodation. To bridge this gap, we propose an AI-driven menstrual detection system that uses closed-circuit (CC) cameras for facial recognition. Leveraging deep learning, our model analyzes facial cues—such as skin texture changes, eye fatigue, and puffiness to detect discomfort non-invasively while preserving privacy. To support this research, we introduce the Menstrual Presence Facial Dataset (MAFD-2024), a curated collection of facial images captured before and during menstruation, annotated for pain indicators. Our hybrid CNN-LSTM model (HCL-MD) enhances detection accuracy by 94.1%, combining spatial (facial features) and temporal (symptom progression) analysis. This system integrates with automated access frameworks, enabling real-time adjustments for affected individuals. Beyond access control, technology can be embedded in telemedicine for remote discomfort assessment or deployed in smart wearables and surveillance systems (e.g., in schools or public transport) to offer timely suggestions. By enabling discreet, real-time support, this AI solution fosters inclusivity and awareness, pioneering the fusion of facial recognition and menstrual health monitoring.