Tend+ AI Safe Shopping Zone

AI-powered personalized shopping platform that protects people with allergies and health conditions. Uses Gemini 3 for chatbot, OCR, and world's first AI ingredient trigger detection.

Comment 1

Inspiration

Every day, millions of people with allergies, chronic conditions, and medication sensitivities struggle to shop safely. Reading ingredient labels is time-consuming, error-prone, and stressful—especially when shopping for products in unfamiliar languages or complex chemical names. We created Tend+ AI Safe Shopping Zone to solve this problem using the power of Google Gemini 3.

What it does

Tend+ AI Safe Shopping Zone is a personalized shopping platform that protects users in real-time:

  • AI Health Chatbot: Captures health information through natural conversation (Gemini 3 NLP)
  • Smart OCR: Scans prescriptions and product labels instantly (Gemini 3 Vision)
  • Real-time Filtering: Automatically filters thousands of products based on individual health profiles
  • AI Ingredient Detective: World's first AI-powered system that discovers hidden trigger ingredients by analyzing "good" vs "bad" products (Gemini 3 Analysis)

Users get instant warnings about allergens, drug interactions, and problematic ingredients—before they buy.

How we built it

Tech Stack:

  • Frontend: Next.js 16, React, TypeScript, Tailwind CSS
  • Backend: Next.js API Routes, Supabase (PostgreSQL)
  • AI: Google Gemini 3 Flash Preview
  • Authentication: Clerk
  • Deployment: Vercel-ready

Gemini 3 Integration:

  1. Natural Language Processing: Health chatbot extracts allergies, medications, and conditions from conversational text
  2. Computer Vision: OCR reads prescriptions and product ingredient labels across multiple languages and formats
  3. AI Analysis: Compares ingredient sets between "good" and "bad" products to identify trigger ingredients with confidence scores

Database Design:

  • PostgreSQL with Row Level Security (RLS) for health data privacy
  • Efficient ingredient matching algorithms for real-time filtering
  • Scalable architecture supporting thousands of products

Challenges we ran into

  • Complex Ingredient Matching: Handling variations in ingredient names across languages (e.g., "Methylparaben" vs "메틸파라벤")
  • OCR Accuracy: Optimizing Gemini 3 Vision for various label designs, fonts, and lighting conditions
  • Performance: Balancing AI response quality with API costs and latency
  • Data Privacy: Implementing secure health data storage with proper RLS policies
  • UX Design: Creating intuitive flows for users to input health data in multiple ways (manual, chat, OCR)

Accomplishments that we're proud of

  • ✅ Built world's first AI-powered ingredient trigger detection system
  • ✅ Achieved high accuracy in multi-language ingredient OCR
  • ✅ Created seamless UX with three input methods (manual, chat, OCR)
  • ✅ Demonstrated with K-Beauty products, proving global scalability
  • ✅ Zero false negatives in allergen detection during testing

What we learned

  • Advanced prompt engineering for healthcare applications
  • Multi-modal AI integration (text + vision + analysis)
  • Real-time filtering algorithms for large product databases
  • Healthcare data privacy and security best practices
  • Importance of user-centric design for sensitive health information

What's next for Tend+ AI Safe Shopping Zone

Short-term:

  • Expand to food, supplements, and personal care categories
  • Add barcode scanning for instant product lookup
  • Enhanced multi-language OCR (currently supports Korean/English labels)

Long-term:

  • Integration with e-commerce platforms (Shopify, WooCommerce)
  • Community features: share safe product lists
  • Medical professional verification system
  • Mobile app (iOS & Android)
  • Partnership with pharmacies and retailers

Vision: Make safe shopping accessible to everyone, everywhere—protecting millions of people with allergies and health conditions worldwide.

Built With

Share this project:

Updates