The Weekly Shop

AI-powered grocery intelligence that turns family receipts into personalised savings opportunities - helping UK families save money on their weekly food shops.

Comment

Inspiration

As a new father facing the UK's relentless cost-of-living crisis, I watched our weekly grocery bills spiral from manageable expenses to budget-breaking shocks - often exceeding £250 per week for our growing family. The frustration wasn't just the rising costs, but the realisation that we were making expensive choices out of habit and brand loyalty without understanding the alternatives. I knew there had to be a smarter way to analyse our spending patterns and discover practical savings opportunities tailored to our family's specific needs. That's when I envisioned The Weekly Shop - an AI-powered assistant that transforms grocery receipts into personalised money-saving strategies, helping families like mine save hundreds of pounds annually while maintaining the quality and nutrition our children deserve.

What it does

The Weekly Shop revolutionises how families approach grocery savings through cutting-edge AI technology.

🤖 Smart Receipt Analysis: Users simply photograph or upload PDF receipts, and our OpenAI Vision-powered OCR instantly extracts every item, price, and detail with remarkable accuracy.

👨‍👩‍👧‍👦 Family-Personalised Recommendations: The app creates detailed family profiles (adults, children's ages, dietary requirements, weekly budgets) and generates contextual suggestions like brand swaps, size optimisations, and healthier alternatives - all tailored to your household's specific needs.

💰 Intelligent Savings Tracking: Real-time progress monitoring shows monthly savings potential, achievement unlocks, and beautiful visualisations of your financial wins.

🎯 Subscription-Based Features: Three tiers (Free, Premium, Family+) offering increasingly sophisticated AI analysis, from 5 basic recommendations to 10 advanced family-optimisation strategies.

📱 Cross-Platform Excellence: Built with Expo/React Native for seamless iOS, Android, and web experiences with offline capabilities and cloud sync.

The result? Families identify an average of £120 in monthly savings opportunities while improving their shopping habits and nutritional choices.

How we built it

The Weekly Shop showcases a sophisticated modern tech stack optimised for performance and scalability.

🚀 Frontend Architecture:

• Expo/React Native for cross-platform mobile and web deployment

• TypeScript throughout for type safety and developer experience

• Custom UI components with beautiful animations and micro-interactions

🧠 AI Integration:

• OpenAI Vision API for advanced OCR and receipt text extraction

• OpenAI GPT-4 for generating contextual, family-specific recommendations

• Custom prompt engineering to deliver British-style, actionable advice

⚙️ Backend & Data:

• Supabase for authentication, real-time database, and cloud storage

• RevenueCat for seamless subscription management, payment processing, and cross-platform monetisation

• Complex state management across subscription tiers and family profiles

• AsyncStorage for offline-first capabilities with cloud sync

📷 Advanced Image Processing:

• Expo ImagePicker & ImageManipulator for receipt optimisation

• Custom image enhancement algorithms for improved OCR accuracy

• Multi-format support (photos, PDFs, various image types)

🏗️ Architecture Highlights:

• Modular service architecture with separate recommendation engines

• Comprehensive error handling and fallback systems

• RevenueCat integration for robust subscription tier management with feature gating

• Achievement system and progress tracking algorithms

Challenges we ran into

1. OCR Accuracy Under Real Conditions: Grocery receipts are notoriously difficult - crumpled paper, poor lighting, faded ink, and varying layouts. We solved this by implementing image optimization pipelines, experimenting with OpenAI Vision prompts, and building robust fallback parsing systems.

2. AI Recommendation Relevance: Generic savings tips are useless. The challenge was making AI understand family dynamics - that a household with toddlers has different needs than one with teenagers. We developed sophisticated prompt engineering and family profile matching algorithms.

3. Cross-Platform Complexity: Building for iOS, Android, and web simultaneously while maintaining native performance was demanding. Expo's limitations with certain APIs required creative workarounds, especially for file handling and image processing.

4. RevenueCat Subscription Integration: Implementing seamless cross-platform subscription management while ensuring proper feature gating, handling edge cases like subscription failures, and maintaining a smooth upgrade/downgrade experience across iOS, Android, and web platforms.

5. Real-Time Data Sync: Ensuring family profiles, receipts, and recommendations stay synchronised across devices while working offline demanded robust conflict resolution and sync strategies.

Accomplishments that we're proud of

🏆 Technical Excellence: Successfully integrated multiple complex APIs (OpenAI Vision, GPT-4, Supabase, RevenueCat) into a cohesive, performant mobile application that handles real-world edge cases gracefully.

🎨 User Experience Design: Created an intuitive, delightful interface that makes complex AI recommendations feel approachable and actionable - users actually enjoy scanning receipts!

💡 AI Innovation: Developed a personalised recommendation engine that considers family size, children's ages, dietary restrictions, and budget constraints to deliver genuinely useful savings advice.

📊 Real Impact: Beta users report identifying £30-70 in monthly savings opportunities, proving the concept's real-world value beyond just technical achievement.

🚀 Production-Ready Architecture: Built with scalability in mind - the codebase is modular, well-documented, and ready for thousands of users without architectural changes.

What we learned

AI Integration Mastery: Gained deep expertise in prompt engineering, vision model optimisation, and building reliable AI-powered features that work consistently in production environments.

Cross-Platform Development: Mastered Expo's capabilities and limitations, learning when to use native modules versus JavaScript solutions for optimal performance.

User-Centered Design: Discovered that successful AI apps require intensive user research - understanding how families actually shop and make decisions was crucial for relevant recommendations.

Real-World Data Challenges: Receipt data is messy and inconsistent. Building robust parsing systems taught us invaluable lessons about handling imperfect real-world inputs.

Subscription Model Psychology: Learned how to design feature tiers that feel valuable rather than restrictive, and implemented RevenueCat for seamless cross-platform monetisation while balancing free functionality with premium value propositions.

Performance at Scale: Optimising image processing, AI API calls, and data sync for mobile devices taught us crucial lessons about building responsive, battery-efficient applications.

What's next for The Weekly Shop

🌟 Enhanced AI Capabilities:

• Integration with grocery store APIs for real-time price comparisons

• Meal planning based on purchased ingredients and family preferences

• Predictive shopping lists that anticipate needs based on consumption patterns

🛒 Shopping Integration:

• Direct integration with major UK supermarket apps (Tesco, ASDA, Sainsbury's)

• Price alerts when recommended products go on sale

• Partnership opportunities with food companies and grocery chains

📱 Social & Gamification:

• Family challenges and savings competitions

• Community sharing of successful swaps and tips

• Leaderboards and advanced achievement systems

🔬 Advanced Analytics:

• Seasonal spending pattern analysis

• Nutritional impact scoring of recommendations

• Carbon footprint tracking for eco-conscious families

🌍 Geographic Expansion:

• Localisation for European markets

• Currency and cultural adaptation for different shopping behaviours

• Partnership opportunities with international grocery chains

The ultimate vision is becoming the essential financial wellness tool for families worldwide - where every grocery shop becomes an opportunity to save money, eat better, and build healthier financial habits.

Testing Details

Login Details

For the best user experience this app has been primarily built as a native iOS app. I have shared the app on testflight with testing@devpost.com, or see public test link https://testflight.apple.com/join/9nmWy1pm

Alternatively the web version is available at https://theweeklyshop.online/

You can login to an existing test account to see pre-populated data such as scanned receipts, items, spending, tips and achievements by using the account below or create your own new account and use the app as it is intended.

Test Email: bolt@test.com Test Password: testing123

Built With

Share this project:

Updates