Meze
A computational food engine for household meal planning — scales recipes to each person’s nutrition goals, aggregates groceries into purchasable units, pushes carts to stores like Kroger, and plans batch prep.
Meze bridges the gap between “what should we eat?” and “dinner is ready” by automating the math of nutrition, the tedium of grocery lists, and the logistics of meal prep. It treats meals as dynamic objects that scale to each eater rather than static recipe text.
Highlights
- Smart meal planning that scales ingredient quantities to per-person calorie/macro targets.
- Intelligent shopping: aggregates ingredients across the week, converts to purchasable units, and pushes matched carts to providers like Kroger.
- Batch-prep optimization via topological sorting of recipe steps, plus AI-generated recipes, images, and metadata.
- Built with Next.js 15, Drizzle ORM on LibSQL/Turso, and NextAuth, deployed to Cloudflare via OpenNext.