Dr. Leo – AI for Mental Wellness

• 

• 

• 

🧠 Inspiration

Mental health struggles are rising among students and young adults, yet access to therapy remains limited. Many people hesitate to seek help due to stigma, cost, or availability. We wanted to build something that could listen without judgment, respond with empathy, and guide users toward emotional balance — available anytime, anywhere. That’s how Dr. Leo, our AI psychiatrist, was born. 🌱

💬 What it does

Dr. Leo is an empathetic AI psychiatrist that engages users in supportive, psychology-aware conversations. It listens, understands emotions, and responds with insights grounded in behavioral science. Each response follows a therapeutic structure: 1️⃣ Explanation of the issue 🧘 Helpful coping strategies or perspectives ❤️ A motivational closing quote

Through this blend of emotional intelligence and artificial intelligence, Dr. Leo makes mental-health guidance accessible, private, and compassionate.

⚙️ How we built it

Frontend: React + Vite interface for smooth, modern chat experience

Backend: FastAPI for asynchronous, lightweight API handling

AI Core: OpenAI-compatible model endpoint from Hugging Face

Knowledge Retrieval: FAISS vector database for context-aware responses (RAG)

Data: Curated mental-health literature and verified psychology resources

Deployment: Render / Vercel for the frontend, and Python-based server for backend

Dr. Leo’s intelligence combines Retrieval-Augmented Generation (RAG) with an empathetic conversation layer that makes replies feel human while staying evidence-based.

🧩 Challenges we ran into

Getting the LLM to stay empathetic and responsible — we had to fine-tune the system prompts carefully.

Handling medical disclaimers and ensuring safety-first responses.

Integrating FAISS search with asynchronous FastAPI routes efficiently.

Time management: balancing technical development with ethical design.

🌟 Accomplishments we’re proud of

Built a functioning AI psychiatrist chatbot with emotional depth.

Designed safe fallback responses for unknown topics.

Delivered a smooth RAG pipeline that retrieves verified knowledge.

Created a user-friendly, compassionate interface that feels truly human.

📚 What we learned

How to blend psychological theory with AI in a meaningful way.

The importance of designing AI that is empathetic, ethical, and human-centered.

Hands-on experience with LLM integration, vector search, and context-aware dialogue systems.

🚀 What’s next for Dr. Leo

Adding voice interaction for hands-free conversations.

Integrating emotion detection from text and tone.

Building a daily mood tracker and journal connected to user sessions.

Partnering with mental-health organizations for real-world pilot testing.

Built With

• fastapi
• hugging
• langcain
• llm
• react