Voice of Help: AI-Driven Mental Health Assessment and Suicide Prevention through Text and Speech

Suicide exists as a major international health issue because it leads to the death of more than $\mathbf{8 0 0, 0 0 0}$ individuals each year. Social media platforms enable people to show emotional distress in the modern digital period thus presenting an essential time for prevention and quick mental health services. Distinct social media posts along with patient questionnaire responses become the training base data for model development. A detection system integrates five classification algorithms which use Logistic Regression, Random Forest, Naive Bayes and Support Vector Machine (SVM) and Artificial Neural Networks (ANN) to optimize different detection tasks. The system displays high accuracy levels through performance assessments conducted with precision, recall, $\mathbf{F 1}$-score calculation and confusion matrix evaluation methods. A web application based on Flask provides secure implementation of the system to users in a user-friendly manner. Users can access the system through text and speech input along with quiz activities and a chatbot helps users receive assistance from the NLP system. The system offers practical solutions to users based on their mental health conditions by recommending mindfulness techniques and critical hotline services. Users gain better transparency and comprehension from the model through visual representation of accuracy graphs and word frequency data. The end-to-end solution was developed to support both mental wellness professionals and online services that help identify vulnerable individuals before they receive proper care for emotional wellbeing and suicide prevention.