Systems & interfaces · 2026
FoodAI: pantry management and recipe ideas
A pantry and recipe app built to help use food before it expires and make deciding what to eat easier.
Local prototype. Source and setup instructions are available on GitHub.
Context
Personal software project · AI-assisted development
My contribution
Ingredient workflows, accounts, persistent data, and recipe-provider integration, developed with AI coding tools.
The takeaway
Use-by estimates and pantry-based recipe suggestions connect ingredient tracking with deciding what to cook.
Why I built it
I got tired of keeping track of when food would expire and deciding what to eat. I hate wasting food, so I built FoodAI with AI coding tools to help me use what needs eating soonest and suggest what to cook.
What it does
- Tracks ingredients in pantry, fridge, or freezer storage, with estimated use-by dates and manual overrides.
- Moves ingredients between the shopping list and pantry, and records when items are used up.
- Suggests recipes from available ingredients, saved dietary preferences, and an optional request, then expands suggestions into recipe cards.
Implementation and scope
The app uses Next.js and TypeScript, account sign-in, and Prisma with PostgreSQL for persistent data. A recipe-provider interface supports OpenAI, Gemini, and a built-in mock fallback. Shelf-life estimates come from an ingredient catalog and fallback rules.
This is a personal prototype. Local verification uses the mock recipe provider; it does not establish the quality of live model-generated recipes.