AI ASSISTANCE FOR MENTAL PEACE

Abstract

Recently, AI-driven journaling tools have emerged as supportive resources for mental health by integrating mood tracking with natural language analysis. Building upon earlier systems such as MindScape, which combines journaling with behavioral information like sleep patterns and location data, and Resonance, which demonstrated that AI-based memory suggestions can enhance mood and reduce depressive symptoms, our application allows users to record daily mood ratings along with written reflections. These entries are analyzed using spaCy-powered techniques including sentiment analysis, entity recognition, and LDA topic modeling to identify recurring themes and emotional patterns. A custom Crisis Detection Model further improves user safety by evaluating keywords, mood variations, and sentiment thresholds to detect potential crisis situations, assess severity, log incidents, trigger in-app alerts, and recommend suitable resources. Developed with Flask and secured through Flask-Login, the system anonymizes journal entries to maintain privacy while functioning as a supportive digital companion that encourages emotional awareness, early detection of distress, and reflective self-care. Planned future enhancements include multilingual support and analysis of long-term mood trends.