Introduction
MealPal is an AI-powered nutrition tracking application with an Oura-inspired glassmorphism design. Users log meals through natural language, get AI-powered nutritional analysis, and chat with Swan — a RAG-powered nutrition assistant.
Tech Stack
Section titled “Tech Stack”| Layer | Technology |
|---|---|
| Frontend | React 19 + TypeScript + Vite |
| Backend | FastAPI + PostgreSQL (RDS) + pgvector |
| Auth | Better-Auth (self-hosted, EdDSA JWTs) |
| AI | OpenAI GPT-4o-mini (free tier), Claude Sonnet (premium) |
| Cache | Valkey (Redis-compatible) |
| Infra | EC2 + RDS + S3 + Amplify (all ca-central-1) |
| Reverse Proxy | Traefik with automatic Let’s Encrypt TLS |
Key Features
Section titled “Key Features”- Natural language meal logging — “I had a chicken caesar salad for lunch”
- AI nutritional analysis — automatic calorie and macro estimation
- RAG chatbot (Swan) — answers nutrition questions using Health Canada, USDA, and WHO data
- Glassmorphism UI — frosted glass design system inspired by Oura Ring
- Demo mode — try the app without signing up
- App: tiny-agents.kovacova.ca
- API: api.tiny-agents.kovacova.ca
- API Docs: api.tiny-agents.kovacova.ca/docs