MINDIFY

MINDIFY is a multimodal stress and emotional wellness assistant built with Gemini 2.0 Flash on Google AI Studio. It combines the power of natural language understanding, speech interpretation, and memory-driven context retrieval to provide personalized stress management and emotional support. Users can express their emotions through text or voice. Gemini interprets the content, detects emotions, and generates empathetic, actionable advice. The system integrates Qdrant vector database to create a personalized memory layer. This memory stores embeddings from over 30 psychology book summaries, CBT (Cognitive Behavioral Therapy) guides, Stoic philosophy texts, and motivational insights. When a user shares their problem, a semantic search retrieves the most relevant passages from Qdrant, grounding Gemini’s response in real psychological frameworks. The AI responds with four key layers: 1️⃣ Empathetic understanding of user emotion 2️⃣ Stress coping suggestions (CBT & mindfulness-based) 3️⃣ Two practical next steps for emotional regulation 4️⃣ Book-based or scientific insight reference MINDIFY architecture fuses multimodal reasoning, emotion detection, and memory retrieval. Voice input adds a deeply human touch, allowing the model to detect mood and tone, while text input enables reflective journaling. Qdrant ensures contextual continuity across sessions, enhancing personalization over time. The project demonstrates how Gemini’s multimodal reasoning + Qdrant’s vector memory can transform emotional wellness applications — making AI not just informative, but emotionally intelligent and healing.