MindChat

About the Project: AI-Powered Mental Health Chatbot

Inspiration

Mental health support is often hard to access, stigmatized, or inconsistent. We wanted to create a tool that could offer compassionate, real-time support using modern AI—something accessible, intelligent, and always available. Inspired by the idea of bridging technology with empathy, we set out to build a mental health chatbot that doesn't just respond, but understands.

How We Built It

We combined several modern technologies to create a seamless, full-stack experience:

• Frontend: Built with React, we designed a clean and intuitive interface that makes interacting with the chatbot easy and engaging.
  
• Backend: Powered by Flask, our backend handles API requests, user sessions, and integrates with our AI engine.
  
• AI & NLP: We use the Gemini API in conjunction with a Retrieval-Augmented Generation (RAG) system to ensure responses are both intelligent and grounded in relevant context.
  
• Database: User conversations and feedback are stored securely using MongoDB Atlas, allowing for scalable and flexible data handling.
  

What We Learned

• How to integrate large language models with custom data via RAG
  
• The importance of tone, UX design, and trust in mental health applications
  
• Real-world implementation of full-stack architecture, balancing performance and privacy
  

Challenges We Faced

• Contextual accuracy: Ensuring the AI doesn’t hallucinate or give poor advice in sensitive contexts
  
• RAG tuning: Integrating relevant sources into the AI’s responses without overwhelming or biasing the conversation
  
• Data privacy: Creating a system that respects user confidentiality while enabling learning and improvement
  

Built With

• flask
• javascript
• mongodb
• mongodbatlas
• python
• rag
• react