The global burden of mental health challenges, particularly among students and young adults, necessitates innovative digital solutions that are both accessible and effective. This paper proposes the design of a web-based Smart Mental Health Assistant to address critical gaps in current digital support tools. The core innovation of the proposed system lies in its integrated architecture, which leverages a fine-tuned Natural Language Processing (NLP) engine for nuanced sentiment analysis and multi-label emotion detection. A key feature is a novel hybrid crisis detection model that combines rule-based filters with contextual machine learning to identify high-risk user communications with high sensitivity. The proposed platform is designed to offer a comprehensive suite of features through a unified web interface, including an empathetic chatbot for real-time interaction, a mood tracking and visualization dashboard, a personalized recommendation engine for coping strategies, and an emergency response module. The design methodology follows an iterative, user-centered approach to ensure usability and engagement. While formal results are pending implementation, the proposed system is designed to surpass existing solutions by integrating advanced NLP capabilities with proactive safety mechanisms, aiming to provide a more holistic, responsive, and secure form of AI- powered mental health support.