Smart Food Recommendation System for Health

Hackathon Project – Confluentia

Problem

People with chronic health conditions such as diabetes, hypertension, and thyroid disorders struggle to know which foods are safe or harmful for their condition. Most nutrition apps are too generic and fail to provide personalized, disease-specific insights. This creates confusion and can even lead to harmful food choices.

Solution

We built a SwiftUI iOS app that leverages Apple’s Foundation Models (Apple Intelligence) to provide personalized food recommendations.

How it works:

1. The user selects or inputs their health condition.
2. The app asks the user to enter a food item they want to eat.
3. The query is passed to Apple’s on-device Foundation Model.
4. The app returns:
   • Recommendation → “Recommended” or “Avoid”
   • Reason → An explanation of why (e.g., glycemic index for diabetes, sodium for hypertension, etc.).
5. Provides personalized nutrition insights, barcode-based food analysis, and dynamically generates custom meal plans based on user-specific health metrics too.

This ensures fast, private, offline and AI-powered insights without sending sensitive data to external servers.

Tech Stack

• Frontend: SwiftUI (iOS 26 Foundation Models support)
• AI: Apple Foundation Models (Foundation + AI frameworks)
• Data Source: Apple’s built-in on device intelligence and prompt engineering

Repo Structure

• HackathonConfluentia → SwiftUI project code
• README.md → Documentation

How to Run

1. Clone or download the repository.
2. Open the HackathonConfluentia.xcodeproj in Xcode 26+ (with iOS 26 SDK).
3. Build and run on the iOS simulator or a physical iPhone running iOS 18/iOS 26.

Demo

• The screenshots show the:
  • Disease input screen
  • Food input screen
  • Output with “Recommended/Avoid” + explanation

AI Tools Used

• Apple Foundation Models (iOS 18 / iOS 26)
• Used for reasoning over food-health queries in natural language.
• Benefits:
  • On-device and privacy-preserving
  • No need for external API calls
  • Optimized for Apple devices

Next Steps

• Expand the system to provide meal planning instead of just single-food queries.
• Integrate with HealthKit to suggest foods based on a user’s biomarker data.
• Add support for multiple languages.