A web app that plans a week of high-protein meals around your calorie and protein targets, with a meal-prep schedule and one shopping list for German supermarkets. Scan your fridge and it shows what you can cook right now.
Hitting a protein goal usually means logging every meal and doing the maths yourself, and most people give up on that within a week.
Pantrio does the planning instead. You set a calorie and protein target once, and it builds a week of simple meals that hit it, using ingredients from Aldi, Lidl, Rewe or Edeka.
Pantrio runs on Next.js 15 with the App Router and TypeScript, hosted on Vercel in Frankfurt. Supabase provides Postgres, sign-in and row-level security, so every user can only read and write their own data.
The week planner, the target calculations and the nutrition maths are pure, tested TypeScript modules. Nutrition values come from Germany's national food database (BLS 4.0) and USDA data files, never from an AI model. Claude is used only to recognise ingredients in typed text and fridge photos, and every AI call has a fallback, so the app keeps working when it fails.
Per-user rate limits with Upstash Redis and a Cloudflare Turnstile check on sign-in protect it from abuse. The app is available in English and German, and CI runs type checks, linting, tests and a dependency audit on every push.
Pantrio is still in development, and the free version is live at pantrio.me. Building it alone taught me to scope tightly: I keep a written plan, ship in small phases and park new ideas until there is evidence they matter.