The Automated Mental Health Classification System classifies users’ mental states—depression, suicidal tendencies, or non-suicidal behavior—based on text or audio input. However, mental health classification faces challenges like data scarcity, symptom variability, and overlapping emotional states. Existing approaches, such as rule-based methods and sentiment analysis models, struggle with contextual nuances, leading to misclassification, while machine learning models require large, labeled datasets that are difficult to obtain due to privacy concerns. To address these issues, the proposed system integrates Logistic Regression, Random Forest, Decision Tree, and Multinomial Naïve Bayes, with Logistic Regression proving the most effective. The Whisper API transcribes speech into text, which undergoes tokenization, stop word removal, lemmatization, and Term Frequency-Inverse Document Frequency feature extraction. Unlike traditional methods, this system not only classifies mental states but also provides real-time, personalized mental health resources, including articles, videos, and helplines. With an intuitive and accessible interface, the system enhances mental health support, offering a technology-driven solution for timely intervention.