At the end of her life, finding meaningful connection with my Grandmother was difficult and the silence between us was painful for both of us. We both sought out connection, but it was often elusive.

It turns out, I was not alone. In America, 12 million caregivers provide unpaid care for those with dementia at home (Alzheimer's Association, 2024). And that number is expected to grow. By 2050, the number of people living with dementia could exceed 152 million (Alzheimer's Association, 2024).

The inspiration for LumiCue came from recognizing the power of a tool that can make personalized encounters with AI could help bridge the communication gap.

LumiCue is an AI-powered app that uses Qloo's Taste AI to help caregivers and those they care for spark meaningful encounters. By combining cultural intelligence with personalized recommendations, LumiCue transforms everyday moments into sparks of connection.

LumiCue consists of two main components working together to deliver culturally intelligent memory care solutions:

React-based user interface for caregivers and memory care patients

The LumiCue frontend is an AI-powered cultural intelligence solution designed to help those with memory loss or trouble communicating connect with their caregivers. The app creates culturally relevant, personalized curated conversation starters that engage the senses and feature music, recipes, photos, and stories tailored to the individual.

Frontend Information: LumiCue Frontend README

Core Features:

• Memory Care Focus: Specifically designed to engage senses and stimulate memories for those with memory loss
• Multi-Modal Content: Combines music (with YouTube integration), recipes, photos, and nostalgia news to allow caregivers to choose which encounters works best
• Feedback System: Collects user preferences to improve future recommendations
• Demo Mode: Showcases the AI pipeline with real-time processing visualization

Technology Stack:

• React 18.2.0
• Tailwind CSS 3.3.0
• Axios 1.6.0
• Lucide React 0.263.1

8-agent AI processing pipeline powered by cultural intelligence

LumiCue Backend is a sophisticated 8-agent AI processing pipeline that powers the cultural intelligence behind LumiCue's personalized memory care recommendations. The API orchestrates multiple AI services to generate culturally relevant, personalized content while maintaining strict data privacy and anonymization protocols.

Backend Information: Backend README

Core Features:

• 8-Agent AI Pipeline: Sophisticated multi-step processing system for content generation
• Cultural Intelligence Integration: Deep integration with Qloo's Taste AI for cultural grounding
• PII Compliance: Complete data anonymization and privacy protection
• Multi-Modal Processing: Handles music, recipes, photos, and story generation
• Real-time Generation: Live content creation with efficient caching
• Safety-First Design: Microwave-only recipes and dementia-appropriate content filtering

Technology Stack:

• FastAPI
• Python 3.9+
• Google Gemini AI (Flash)
• Qloo Taste API
• Google Vision AI
• YouTube Data API

8-Agent Pipeline:

1. Information Consolidator - Data anonymization and theme selection
2. Simple Photo Analysis - Image processing with Google Vision AI
3. Qloo Cultural Intelligence - Cultural preference grounding used for content creation
4. **Music Curation ** - Creative Commons music discovery
5. **Recipe Selection ** - Safe, culturally relevant recipe curation
6. **Photo Description ** - Culturally relevant photo storytelling
7. Nostalgia News Generator - Personalized historical storytelling
8. Dashboard Synthesizer - Final content assembly and formatting

• Node.js 18+ installed
• Python 3.9+ installed
• npm or yarn package manager
• pip package manager
• API keys for Qloo, Google Gemini, Google Vision, and YouTube Data API

# Clone the frontend repository
git clone https://github.com/your-username/lumicue-frontend.git
cd lumicue-frontend

# Navigate to the frontend directory
cd frontend

# Install dependencies
npm install

# Start development server
npm start

The frontend application will launch at http://localhost:3000

# Development server
npm start

# Build for production
npm run build

# Run tests
npm test

# Eject configuration (not recommended)
npm run eject

# Build optimized production bundle
npm run build

# The build folder contains the production-ready application
# Deploy the contents to your web hosting service

# Clone the backend repository
git clone https://github.com/your-username/lumicue-backend.git
cd lumicue-backend

# Navigate to the backend directory
cd backend

# Create virtual environment
python -m venv venv

# Activate virtual environment
# On Windows:
venv\Scripts\activate
# On macOS/Linux:
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Set up environment variables
cp .env.example .env
# Edit .env with your API keys

# Start development server
uvicorn main:app --reload --host 0.0.0.0 --port 8000

The backend API will be available at http://localhost:8000

# Development server with hot reload
uvicorn main:app --reload

# Production server
uvicorn main:app --host 0.0.0.0 --port 8000

# Run tests
pytest

# Check code formatting
black .

# Lint code
flake8 .

Create a .env file with the following variables:

# API Keys
QLOO_API_KEY=your_qloo_api_key_here
GOOGLE_API_KEY=your_google_api_key_here
YOUTUBE_API_KEY=your_youtube_api_key_here

# Environment
ENVIRONMENT=development
DEBUG=True

# CORS Settings
ALLOWED_ORIGINS=http://localhost:3000,https://your-frontend-domain.com

1. Start the backend (Terminal 1):

cd lumicue-backend/backend
source venv/bin/activate  # or venv\Scripts\activate on Windows
uvicorn main:app --reload --host 0.0.0.0 --port 8000

2. Start the frontend (Terminal 2):

cd lumicue-frontend/frontend
npm start

3. Access the application:
   • Frontend: http://localhost:3000
   • Backend API: http://localhost:8000
   • API Status: http://localhost:8000/api/status

The frontend automatically connects to the backend through the following endpoints:

• POST /api/dashboard - Generate personalized dashboard content
• GET /api/status - Get system health and agent status

LumiCue integrates with multiple AI services. You'll need API keys for:

1. Qloo Taste AI - Cultural intelligence and recommendations
2. Google Gemini AI - Content generation and storytelling
3. Google Vision AI - Image analysis and cultural context
4. YouTube Data API - Creative Commons music discovery

• Simple Interface: Clean, intuitive design prioritizing accessibility
• Cultural Intelligence: Content that respects and celebrates diverse backgrounds
• Safety-First: All recommendations prioritize patient safety and dignity
• Conversation Starters: Guided prompts to facilitate meaningful interactions

• Multi-Sensory Engagement: Music, recipes, photos, and stories
• Culturally Relevant: Content tailored to personal heritage and preferences
• Memory-Safe: Appropriate complexity and familiar references
• Personalized: AI-driven recommendations that adapt over time

• 8-Agent AI Pipeline: Sophisticated orchestration of multiple AI services
• Real-time Processing: Live demonstration of AI agent execution
• Bulletproof Fallbacks: System works even when APIs are unavailable
• Privacy-First: Complete data anonymization and PII compliance
• Production-Ready: Deployed on Google Cloud with enterprise-grade reliability

Alzheimer's Association. (2024). 2024 Alzheimer's Disease Facts and Figures. Retrieved from https://www.alz.org/alzheimers-dementia/facts-figures