Caltrax

Premium nutrition & health tracking. Next.js 14 (App Router) + TypeScript (strict) + Tailwind + Prisma/PostgreSQL.

Getting started

npm install
cp .env.example .env        # fill in DATABASE_URL
npx prisma migrate dev --name init
npm run dev

Open http://localhost:3000 — it redirects to /dashboard, which renders live against the goal engine (see below) using a sample profile until auth + real data wiring lands.

What’s built (Phase 1 foundation)

• Project scaffold: Next.js + TypeScript (strict mode, no implicit any, no unchecked indexed access) + Tailwind, ready to npm install and run.
• Design system (tailwind.config.ts, globals.css): near-black dark-mode-first palette, glass card primitive (24px blur, 22px radius per spec), semantic accent colors (blue/green/orange/red mapped to info/success/warning/danger), macro-specific hues, reduced-motion support, visible focus rings.
• Goal engine (src/lib/goalEngine.ts): real, tested-by-hand math — Mifflin-St Jeor BMR, activity-multiplier TDEE, goal-specific protein/fat/carb splits, a hard safety floor on minimum calories, and rate-of-change capped at 1% bodyweight/week regardless of user input. Pure functions, no UI coupling.
• Database schema (prisma/schema.prisma): Users, Profiles, Foods (multi-source: Open Food Facts / USDA / custom), MealEntries, WeightLogs, WaterLogs, Reminders.
• Dashboard (src/app/dashboard/page.tsx + src/components/dashboard/*): Calories-remaining hero ring, macro rings (protein/carbs/fat/fibre), hydration widget with quick-add, weight trend chart (Recharts) — all live components, not static mockups, wired to calculateGoals().
• Supplement education data (src/lib/supplements.ts): the 12 supplements listed in the spec, each with a general evidence-based range, timing, interactions and a source link — paired with a mandatory disclaimer constant that must render alongside every card. This is reference information, not personalized dosing.

Roadmap (not yet built — in priority order)

Phase 1 — finish the core loop

1. Auth (email/password + session, OAuth-ready) and onboarding flow that collects the UserProfileInput fields and calls calculateGoals() to create the Profile row.
2. Food search + barcode lookup against Open Food Facts (primary) and USDA FoodData Central (secondary), normalized into the Food model.
3. Meal logging UI (breakfast/lunch/dinner/snack) writing MealEntry rows, with the dashboard reading real aggregates instead of the sample data.
4. Weight/water logging UI writing to WeightLog/WaterLog.

Phase 2 — camera & AI

1. Browser camera access (getUserMedia) for barcode/QR scanning (a zxing-wasm dependency is already in package.json for this) and nutrition-label OCR.
2. AI meal photo recognition + AI daily/weekly/monthly insights, calling the Anthropic API server-side with structured JSON output.

Phase 3 — depth

1. Recipes, meal templates, weekly meal planner, favourites/recent foods.
2. Full micronutrient tracking (vitamins A/B/C/D/E/K, minerals, omega-3/6) end to end through logging, storage and dashboard charts.
3. Reminders (meal/water/exercise/weight-check/supplement) with push notifications.
4. Achievements/badges, weekly/monthly/yearly reports.

Phase 4 — platform

1. PWA packaging (manifest + service worker + offline cache).
2. Wearable/health-platform integrations (Apple Health, Health Connect, Garmin, Fitbit, Whoop, Oura, smart scales, CGMs).
3. Premium subscription, social features, family accounts, multi-language.

Notes on scope

This spec describes a product on the scale of MyFitnessPal/MacroFactor — a realistic build is months of engineering, not one pass. This foundation is real, working code (not mockups) for the pieces most projects get wrong if rushed: the math, the schema, and the visual system. Everything above builds directly on top of it without needing rework.