Gainz Trackerz
GPT-4 nutrition + fitness tracker
Runs locally via Docker Compose; source on GitHub.
Gainz Trackerz is a full-stack fitness/nutrition app as a Turborepo monorepo: a Next.js 14 + TypeScript web client over a FastAPI gateway (async SQLAlchemy + Postgres + Redis, JWT) plus food, ml, and workout microservices in Docker Compose (with MinIO + pgAdmin). The genuinely AI part is an OpenAI GPT-4 pipeline in the gateway that parses natural-language meal descriptions into structured items, enriches them against USDA FoodData Central, Nutritionix, and Spoonacular, and falls back to a second GPT-4 estimate when sources disagree. (The photo/voice recognition endpoints are scaffolded stubs, not trained models.)
- Next.js 14
- TypeScript
- FastAPI
- Python
- OpenAI GPT-4
- PostgreSQL
- Redis
- Docker
- Turborepo
Architecture · free-text meal → validated nutrition
Natural-language meal input
A free-text description (“two eggs and a slice of toast”) enters the FastAPI gateway.
GPT-4 parsing
OpenAI GPT-4 parses it into structured food items with quantities.
Multi-source enrichment
Each item is enriched against USDA FoodData Central, Nutritionix, and Spoonacular.
Disagreement fallback
When the sources disagree, a second GPT-4 estimate breaks the tie.
Structured + tracked
Validated items persist to Postgres and are tracked across the monorepo's services.
- Services
- 3 + FastAPI gateway
- Nutrition sources
- USDA · Nutritionix · Spoonacular
- AI
- GPT-4 meal parsing
- Dev stack
- Turborepo · Docker (8 services)