Smart Mental Health Companion

Initial full-stack website scaffold for the AI-assisted mental health support project.

Stack

• Frontend: React + Vite
• Backend: Express + MongoDB Atlas
• Current AI layer: Python emotion-analysis microservice with JavaScript fallback
• Online chatbot layer: Gemini API support with local fallback when no key or quota is available
• Optional local chatbot layer: Ollama local model support

Run the project

Backend

1. Copy backend/.env.example to backend/.env
2. Add your MongoDB Atlas MONGO_URI
3. Add a Gemini API key for online Dr. Beep replies:

GEMINI_API_KEY=your_gemini_api_key
GEMINI_MODEL=gemini-2.5-flash

4. Optionally run Ollama locally for offline advanced Dr. Beep replies:

Recommended local model:

ollama run llama3.1:8b

5. Start the Python AI service:

& "C:\Users\sambh\AppData\Local\Programs\Python\Python312\python.exe" "d:\vscode\AIChatbot\ai_service\server.py"

6. Run:

cd backend
npm install
npm run dev

Frontend

1. Copy frontend/.env.example to frontend/.env
2. Run:

cd frontend
npm install
npm run dev

Current project modules

• AI chatbot interaction area
• Python-backed emotion analysis service
• Optional Gemini-powered online conversational replies
• Optional Ollama-powered advanced conversational replies
• Personalized wellness suggestion flow
• Mood tracking history panel
• Atlas-ready data storage structure

Next recommended upgrades

• Stronger Python NLP model or richer local model orchestration
• User authentication
• Better mood analytics dashboard
• Crisis support and safety escalation messaging