Inspiration
The global aging population faces a growing challenge: loneliness and disconnection from caregivers. We saw family members struggling to stay connected with elderly loved ones while balancing busy lives, and seniors feeling isolated between caregiver visits. We asked ourselves: What if we could create a companion that's always there, one that brings joy, encourages healthy habits, and keeps families connected?
Inspired by the emotional bond between pets and their owners (think Nintendogs!), we envisioned BuddyCare - a virtual companion puppy named Buddy who provides daily emotional support to elderly users while keeping caregivers informed about their wellbeing in real-time.
What it does
BuddyCare is a dual-interface care platform that connects elderly users with their caregivers through an AI-powered virtual companion.
For Elderly Users:
- Meet Buddy - a beautiful 3D interactive golden retriever puppy that responds to their daily activities with 5 different moods (sleeping, awake, smiling, excited, love)
- Voice-enabled AI Chat - talk to Buddy using speech recognition; Buddy responds with text-to-speech, making conversations natural and accessible
- Daily Missions - gentle reminders for medicine, meals, exercise, and mood check-ins with satisfying completion celebrations
- Heart Button - send love to caregivers with a single tap, creating emotional connection across distance
For Caregivers:
- Real-time Dashboard - monitor mission completions, step counts, and puppy status live
- AI-Generated Daily Summaries - intelligent analysis of conversations with automatic concern detection (pain, loneliness, anxiety)
- Emotion Tracking - 7-day emotional trend visualization to spot patterns
- Heart Notifications - receive love hearts from elderly users with instant toast notifications
- Reward System - earn coffee coupons after collecting 20 hearts (gamified appreciation!)
How we built it
Frontend:
- React 18 + TypeScript for type-safe, component-based architecture
- Three.js + React Three Fiber for the 3D Nintendogs-style puppy with real-time animations (breathing, tail wagging, mood effects like floating hearts and sparkles)
- Framer Motion for buttery-smooth animations throughout the app
- Tailwind CSS + shadcn/ui for a modern, accessible UI with large touch targets optimized for elderly users
Backend (Lovable Cloud):
- Supabase Realtime for instant synchronization between elderly and caregiver dashboards
- Edge Functions for AI-powered features:
ai-chat: Handles conversation with Buddy, analyzes user emotions (happy, sad, anxious, pain_suspected, neutral)summarize-daily-chat: Generates intelligent daily summaries with concern detection
- Lovable AI Gateway (Google Gemini) for natural language processing
- PostgreSQL with Row Level Security for secure multi-user data isolation
Key Database Tables:
profiles(user management with elderly-caregiver linking)conversations(chat history with emotion tags)mission_completions(daily activity tracking)hearts(love sent between users)daily_summaries(AI-generated summaries cached)caregiver_coupons(reward system)
Challenges we ran into
3D Puppy Positioning - Getting the mood effects (floating hearts, sparkles, Z's for sleeping) to display correctly without being cut off required careful camera positioning and disabling rotation controls
Real-time Sync Complexity - Ensuring mission completions, hearts, and conversations sync instantly between elderly and caregiver views required careful Supabase Realtime channel management
Emotion Detection Accuracy - Training the AI to understand context (e.g., "I'm not alone" shouldn't trigger loneliness alerts) required extensive prompt engineering with explicit negation handling
Elderly-Friendly UX - Designing for seniors meant larger fonts (
text-elderly-lg), simplified navigation, voice-first input, and high-contrast visual feedbackMobile Dialog Centering - Ensuring popups appear centered on phone screens required custom CSS with
calc(100% - 2rem)width calculations and proper transform translations
Accomplishments that we're proud of
- The 3D Puppy - Buddy is absolutely adorable! Built entirely with Three.js primitives (no external models), with 5 distinct moods and smooth animations that genuinely make you smile
- Voice Interaction - Full speech-to-text and text-to-speech support makes the app accessible to users who struggle with typing
- Intelligent Concern Detection - The AI accurately identifies when elderly users mention pain, loneliness, or anxiety, alerting caregivers without false positives
- Gamified Caregiving - The heart-to-coupon reward system turns caregiving into a positive feedback loop
- Pixel-Perfect Mobile Experience - Every component is optimized for elderly users on mobile devices
What we learned
- Accessibility is Design - Building for elderly users forced us to make better design decisions that improve the experience for everyone
- AI Prompt Engineering - Creating reliable emotion detection and summary generation required iterative prompt refinement
- Real-time Architecture - Supabase Realtime channels need careful subscription management to avoid memory leaks
- 3D on the Web - React Three Fiber makes 3D surprisingly approachable, but performance optimization is crucial for mobile
What's next for BuddyCare
- Step Counter Integration - Connect with wearables (Apple Watch, Fitbit) for automatic step tracking
- Emergency SOS - One-tap emergency call feature with automatic location sharing
- Multiple Elderly Support - Allow caregivers to monitor several elderly users
- Customizable Missions - Let caregivers set personalized daily goals
- Video Calling - Built-in video chat between elderly users and caregivers
- Health Metrics Dashboard - Integrate blood pressure, heart rate, and medication reminders
- Buddy's Growth - Let the puppy "grow" and unlock new animations based on mission streaks
Your caring companion, every step of the way. 🐕💕
Built With
- postgresql
- react
- supabase
- three.js
Log in or sign up for Devpost to join the conversation.