macromaxxing

A recipe nutrition tracker and workout logger for fitness enthusiasts who meal prep. Track macros per portion, plan weekly meals, and log workouts with biomechanical validation.

Features

• Recipe Management — Create recipes with ingredients, track cooked weight vs raw weight, record fat discarded during cooking, define portion sizes. Supports subrecipes (recipes as components of other recipes)
• Recipe Import — Import recipes from URLs (JSON-LD structured data → AI fallback) or pasted text
• Premade Meals — Add pre-made products via nutrition label data or URL parsing
• Macro Visualization — MacroRing (donut chart showing P/C/F caloric ratio), MacroBar (stacked horizontal bar), per-portion readouts
• Ingredient Units — AI-populated unit conversions (tbsp, scoop, pcs) with gram equivalents; density-based volume calculations
• Preparation Tracking — Auto-extracts preparation descriptors ("minced", "finely chopped") from ingredient names
• AI-Powered Ingredient Lookup — USDA FoodData Central API priority, AI fallback (Gemini/OpenAI/Anthropic BYOK). Batch lookups and model fallback configurable per user
• Weekly Meal Planning — Template-based weekly planner with per-plan recipe inventory and portion tracking
• Workout Tracking — Template-based training with checklist-driven session logging, supersets, auto-generated warmup/backoff sets, fatigue-tier-based rest timers, and interactive body map
• Rest Alerts — Optional Web Push notifications when a rest ends, including after the installed app is closed or the phone is locked, with local-only offline fallback
  • Enable them under Settings → Rest alerts, approve browser notifications, then use Test notification to verify delivery on the device.
  • Delivery is intentionally best-effort: Queue retries are disabled, avoiding stale late alerts at the cost of a rare missed accepted alert if consumer processing fails.
  • Pages previews disable rest alert mutations so preview traffic cannot create production notification jobs.
• Training Locations — Define the places you train (gym, home, hotel) with equipment checklists; templates and sessions warn when an exercise needs equipment the location doesn't have
• Auth — Google/GitHub OAuth via Clerk
• Responsive Design — Two-column editor on desktop, mobile-first with bottom tab navigation

Tech Stack

Frontend:

• React 19 + TypeScript
• Vite 7 + Tailwind CSS 4
• tRPC + TanStack Query for data fetching
• react-router-dom for routing
• @dnd-kit for drag-and-drop (ingredient reordering, meal plan allocation)

Backend:

• Cloudflare Pages Functions (Hono + tRPC)
• Cloudflare D1 (SQLite) with Drizzle ORM
• Cloudflare Queues + a dedicated Worker for delayed rest-alert delivery
• Clerk for authentication (cookie-based sessions)
• AES-GCM encrypted API key storage

Project Structure

src/
├── components/
│   ├── layout/              # RootLayout, Nav (desktop top + mobile bottom tabs + RestTimer)
│   └── ui/                  # Button, Card, Input, NumberInput, Select, Switch, Spinner, etc.
├── features/
│   ├── recipes/
│   │   ├── components/      # MacroRing, MacroBar, MacroReadout, PortionPanel, RecipeImportDialog,
│   │   │                    #   PremadeDialog, SubrecipeExpandedRows, IngredientSearchInput, etc.
│   │   ├── hooks/           # useRecipeCalculations
│   │   ├── utils/           # macros.ts, format.ts
│   │   ├── RecipeListPage.tsx
│   │   └── RecipeEditorPage.tsx
│   ├── ingredients/
│   │   ├── components/      # IngredientForm
│   │   └── IngredientListPage.tsx
│   ├── mealPlans/
│   │   ├── components/      # InventorySidebar, WeekGrid, DayColumn, MealCard, DayTotals,
│   │   │                    #   WeeklyAverages, SlotPickerPopover, etc.
│   │   ├── MealPlanListPage.tsx
│   │   └── MealPlannerPage.tsx
│   ├── workouts/
│   │   ├── components/      # BodyMap, ExerciseSetForm, SetRow, SupersetForm, SessionReview,
│   │   │                    #   TimerMode, TimerRing, RestTimer, RestAlertsSection, etc.
│   │   ├── hooks/           # Rest-alert delivery controller and session hooks
│   │   ├── store/           # Persisted workout session/timer state
│   │   ├── utils/           # formulas.ts, sets.ts, export.ts
│   │   ├── RestTimerContext.tsx
│   │   ├── WorkoutListPage.tsx
│   │   ├── WorkoutTemplatePage.tsx
│   │   └── WorkoutSessionPage.tsx
│   └── settings/
│       └── SettingsPage.tsx
├── lib/
│   ├── trpc.ts              # tRPC client setup
│   ├── user.tsx             # useUser() hook (Clerk)
│   └── cn.ts                # clsx + tailwind-merge utility
└── index.css                # Design tokens (surfaces, ink, macro colors, etc.)

packages/db/                 # Shared package @macromaxxing/db
├── schema.ts                # All tables, including pushSubscriptions and restNotificationJobs
├── relations.ts             # Drizzle relations
├── types.ts                 # Inferred types
├── custom-types.ts          # TypeID helpers, AiProvider, FatigueTier, MuscleGroup, SetMode, etc.
└── preparation.ts           # Preparation descriptor extraction

workers/functions/
├── api/[[route]].ts         # Hono entry: Clerk auth middleware → tRPC handler
└── lib/
    ├── router.ts            # tRPC app router (recipe, ingredient, settings, ai, dashboard, mealPlan,
    │                        #   user, workout, analytics, restNotifications)
    ├── trpc.ts              # tRPC context + procedures
    ├── auth.ts              # Clerk cookie verification
    ├── db.ts                # Drizzle D1 setup
    ├── ai-utils.ts          # Multi-provider AI client, model fallback, JSON-LD extraction
    ├── crypto.ts            # AES-GCM encryption helpers
    ├── constants.ts         # Shared constants + Zod schemas
    ├── utils.ts             # toStartCase, extractPreparation, etc.
    └── routes/
        ├── recipes.ts       # CRUD + ingredients + subrecipes + premade meals
        ├── ingredients.ts   # CRUD + findOrCreate + batchFindOrCreate + units
        ├── mealPlans.ts     # CRUD + inventory + slot allocation
        ├── workouts.ts      # Exercises, templates, sessions, sets, muscle stats, import
        ├── ai.ts            # lookup, estimateCookedWeight, parseRecipe, parseProduct
        ├── settings.ts      # AI config + body profile
        ├── restNotifications.ts # Push subscription, scheduling, cancellation, and test notification API
        └── user.ts          # User endpoints

workers/rest-notifications/  # Queue consumer that claims due jobs and sends Web Push

scripts/
└── seed-exercises.ts        # System exercises with muscle group mappings + strength standards

AI Features

Nutrition lookup priority: USDA FoodData Central API → AI (user's configured provider)

Provider	Default Model	Fallback Chain
Gemini	gemini-2.5-flash	→ gemini-2.5-flash-lite-preview → gemma-3-27b-it
OpenAI	gpt-4o-mini	—
Anthropic	claude-sonnet-4-20250514	—

Capabilities:

• Ingredient nutritional data lookup (macros, density, units per 100g raw)
• Recipe parsing from URLs (JSON-LD structured data) or text (AI)
• Product nutrition parsing from URLs (JSON-LD Product → AI fallback)
• Cooked weight estimation from ingredients + instructions
• Batch ingredient lookups (single AI call for multiple ingredients)
• Model fallback on rate limits (429 → next model in chain)

Per-user settings (both off by default):

• batchLookups — batch N ingredient AI calls into 1
• modelFallback — retry with cheaper models on 429

Design System

Dark theme with warm soapstone undertones. Key tokens in src/index.css:

• Surfaces: surface-0 (base), surface-1 (cards), surface-2 (hover/elevated)
• Ink: ink (primary text), ink-muted (secondary), ink-faint (tertiary)
• Macro colors: macro-protein (copper), macro-carbs (golden), macro-fat (olive), macro-kcal (warm orange), macro-fiber (sage)
• Accent: Copper (oklch(0.72 0.15 50))
• Depth: Borders-only strategy (no shadows), edge border color
• Radius: Sharp (4px/6px) for instrument-grade precision

Development

# Install dependencies
yarn

# Run dev server (frontend + backend)
yarn dev

# Build
yarn build

# Lint/format
yarn fix

# Database migrations
yarn db:generate   # Generate migration from schema changes
yarn db:migrate    # Apply migrations to local D1

# Run tests
yarn test

Environment Variables

Frontend (.env.local):

• VITE_CLERK_PUBLISHABLE_KEY — Clerk publishable key

Workers (.dev.vars locally, Cloudflare dashboard for production):

• ENCRYPTION_SECRET — 32-byte hex string for AES-GCM key encryption
• USDA_API_KEY — USDA FoodData Central API key
• CLERK_PUBLISHABLE_KEY — Clerk publishable key
• CLERK_SECRET_KEY — Clerk secret key
• VAPID_PUBLIC_KEY — URL-safe public key used for browser push subscriptions
• VAPID_PRIVATE_KEY — matching private key used only by server-side delivery
• REST_ALERTS_ENABLED — runtime feature gate; production enables it while Pages previews set it to false in workers/wrangler.toml

Rest alerts use the macromaxxing-rest-notifications Cloudflare Queue. Production needs the same VAPID key pair in the Pages project and the dedicated macromaxxing-rest-notifications Worker. The deploy workflow creates the queue, deploys its consumer before Pages, and configures both runtimes from the VAPID_PUBLIC_KEY and VAPID_PRIVATE_KEY GitHub Actions secrets. Never add either key to wrangler.toml.

License

MIT