ProjectsMarch 15, 2025

Mohamed Elbarry
image

Technologies used

React Native
Expo
TypeScript
Supabase
PostgreSQL
Node.js

Overview

cAIlories is a cross-platform (iOS, Android, Web) health and nutrition app that replaces manual calorie logging with a "point, snap, log" flow: the user takes a photo of their plate, and the app uses AI to identify the food, estimate portions, and log calories and macros in seconds. I built it with an in-app AI nutrition coach, smart dashboards, weight tracking (including Apple HealthKit sync), gamification (achievements, streaks, leaderboard), and a freemium subscription model so premium features like unlimited AI scans and the coach are enforced server-side.

Key Features

  • AI photo recognition: Single-photo meal logging; the app identifies food, estimates portions, and logs calories and macros (protein, carbs, fat, fiber).
  • AI nutrition coach: Chat interface for meal suggestions, goal tracking, and motivation; conversation history stored and contextual to the user.
  • Smart dashboards: Daily and weekly progress, macro breakdowns, TDEE-based calorie targets, and fasting windows.
  • Weight tracking: Manual entries and Apple HealthKit sync (read/write weight).
  • Gamification: 27 achievements, streaks, points, and leaderboard.
  • Weekly recap: Story-style recap of the week's nutrition.
  • iOS home screen widget: Native Swift widget (Expo native target) for at-a-glance progress.
  • Monetization: Freemium with cAIlories Pro (RevenueCat monthly/annual); paywall and server-side subscription checks so premium features are enforced in the backend.
  • Localization: 12 languages (en, es, fr, de, it, nl, ar, ru, sv, ja, ko, zh-Hans) with react-i18next.

Technologies Used

  • Frontend / mobile: React Native with Expo (SDK 54), TypeScript, Expo Router, NativeWind (Tailwind-style), TanStack Query, Zustand, expo-camera/expo-image-picker, expo-notifications, expo-tracking-transparency, react-native-gifted-charts.
  • Backend: Supabase (Auth, PostgreSQL, RLS on all tables, Storage for food images), Edge Functions (Deno) for analyze-food, chat-with-coach, revenuecat-webhook, delete-account, get-analytics.
  • Integrations: RevenueCat (in-app purchases, Premium entitlement, webhook for server sync), Apple HealthKit (weight), PostHog (analytics), TikTok Ads / SKAdNetwork (attribution), native iOS widget via Expo @bacons/apple-targets and App Groups.
  • Tooling: Zod for validation, ESLint, Prettier.

Challenges and Learnings

All sensitive Edge Functions validate the JWT and check profiles.is_pro and subscription_expires_at from the database so premium logic is server-authoritative rather than client-only. The AI flow is two-stage—"identify" (food name and description from the image) then "calculate" (nutrition from description and optional portion)—which kept the pipeline clear and supported multiple languages for accessibility. Designing RLS so every table is scoped by auth.uid() (or explicit policies) ensured no client-side trust for subscriptions or PII. The onboarding flow is a structured multi-step sequence (goal, weight, target weight, height, age, activity, profile, feature previews, commitment, notification permission, summary) with progress and analytics events; keeping it maintainable required clear state (Zustand) and step boundaries.

Outcome

I shipped a production-oriented cross-platform nutrition app with AI-powered meal logging, an in-app coach, and subscription monetization. The backend stays in sync with RevenueCat via webhook, and premium features are enforced in Edge Functions and with trigger-protected profile columns. The app supports 12 languages, integrates HealthKit and a native iOS widget, and uses PostHog and SKAdNetwork for analytics and attribution. The combination of RLS, server-side subscription checks, and a two-step AI pipeline keeps the product secure and scalable as usage grows.

High-Performance Agency Website

A modern, high-performance agency website built using Next.js 16, React 19, and Tailwind CSS v4.
Read case study
HireUp - Freelance Marketplace

A high-performance, full-stack freelance marketplace featuring real-time messaging, secure payments, and role-specific dashboards.
Read case study
Construction Company - Private Client Media Management

Developed a private web and mobile app for a construction company to help manage on-site media, allowing workers to capture and upload photos and videos with location assignment and access controls.
Read case study
Lumin Search - AI-Powered E-commerce Search Engine

Developed and integrated AI-powered search solutions for e-commerce platforms, focusing on semantic search features and real-time catalog syncing.
Read case study
Lumin AI - E-commerce Chatbot Platform

Developed and integrated AI solutions for e-commerce chatbot platform, including context-aware conversation flows and multilingual support.
Read case study