🐾 DesktopPet – Your Intelligent Focus Companion
Inspiration
As students, we’ve experienced both extremes of productivity:
- Hyperfocus — working for hours without drinking water, moving, or taking breaks.
- Mental fatigue — staring at a screen, unable to focus, stressed but pushing through anyway.
Productivity culture often ignores basic physical and emotional needs. We wanted to design something that quietly protects them.
What if your desk companion could care about your well-being while you work?
That question became DesktopPet.
What It Does
DesktopPet is an intelligent desktop companion that adapts to your work behavior in real time.
💧 Adaptive Hydration Management
- Tracks hydration schedules in a database
- Updates reminders dynamically based on the last drink timestamp
- Reduces intervals if reminders are ignored
- Makes reminders less frequent once healthy habits are formed
- Adapts to behavior instead of using rigid timers
🧠 Stress Detection & Focus Monitoring
Using a live camera feed and computer vision models, DesktopPet:
- Detects eye fatigue
- Recognizes facial tension
- Identifies prolonged work sessions
- Monitors signs of stress and attentional fatigue
When stress patterns are detected, it responds intelligently.
🌿 Real-Time De-Stressing Assistance
- Suggests short breathing exercises
- Encourages micro-breaks
- Generates supportive voice prompts
- Uses conversational AI for emotional check-ins
DesktopPet isn’t just a timer.
It’s a behavior-aware system that learns and responds.
How We Built It
🌐 Web Application
Frontend
- React
- TypeScript
- Designed in Figma
Backend
- Flask
- RESTful API architecture
- MongoDB for persistent user data and hydration logs
🤖 Firmware & Embedded System
DesktopPet is powered by custom firmware running on embedded hardware:
- ESP32 — handles device logic, Wi-Fi communication, and backend synchronization
- ESP32-CAM — captures live camera feed for stress and fatigue detection
- LCD Screen — displays hydration reminders, emotional states, and interactive animations
- Speaker — delivers AI-generated voice prompts and reminders
- Microphone — enables voice interaction and conversational check-ins
The ESP32 communicates with the Flask backend over Wi-Fi, receiving adaptive reminder schedules and sending behavioral data in real time.
This hardware-software integration transforms DesktopPet from a web app into a physical, interactive desk companion.
🧠 Intelligence Layer
- Google Gemini API — adaptive AI responses and conversational logic
- ElevenLabs API — realistic voice synthesis for reminders and support
- Presage —facial analysis for stress and fatigue detection
- OpenCV — processes ESP32 camera footage
Infrastructure
- MongoDB — hydration logs, behavior tracking, timestamps
- Dynamic backend scheduling logic
- Real-time camera processing pipeline
- Embedded-to-server communication over REST endpoints
Challenges We Ran Into
- Designing adaptive reminder logic instead of fixed timers
- Integrating live camera analysis with backend decision-making
- Managing hydration timestamps across time zones
- Preventing false positives in stress detection
- Optimizing ESP32 performance for real-time interaction
- Synchronizing firmware and backend communication reliably
Accomplishments We're Proud Of
- Building a behavior-adaptive hydration system
- Creating a real-time stress detection loop
- Successfully integrating hardware, firmware, AI, and backend systems
- Designing something that genuinely improves daily student life
What We Learned
- Adaptive systems are more powerful than static timers
- Behavioral design matters as much as technical implementation
- Hardware-software integration requires careful system architecture
- AI is most impactful when it enhances everyday habits
What's Next
- Personalized stress profiling
- Gamified habit-building
- Long-term wellness analytics dashboard
- More advanced on-device processing
- Campus pilot testing with students
Built With
- React
- TypeScript
- Flask
- MongoDB
- OpenCV
- Google Gemini API
- ElevenLabs API
- Presage
- ESP32
- ESP32-CAM
- LCD Display
- Speaker Module
- Microphone Module
- Figma
Log in or sign up for Devpost to join the conversation.