Inspiration
A device or method to track focus time during work or study sessions, remind user when they aren't focused or they are slouching and reward them for every minute they focus with Solana tokens.
What it does
Snitch_It is an innovative productivity enforcement system that combines:
Computer Vision (Python + MediaPipe) for real-time posture tracking IoT Hardware (Arduino + Grove Shield) for physical presence detection Web3/Solana Integration (DePIN) for token-based rewards
The system monitors whether you're actually working while sitting at your desk with proper posture. When you complete a focused work session successfully, you earn Solana tokens – literally getting paid for staying focused!
How it was built
Snitch_It System Flow
THE BRAIN (Laptop/Python): Captures webcam feed in real-time Uses MediaPipe for skeleton/posture tracking Detects: slouching, head tilt, phone distraction Sends status to Arduino via Serial
THE BODY (Arduino/Hardware): Ultrasonic sensor checks desk proximit RGB LCD displays focus status & earnings Buzzer alerts on posture violations Syncs focus state with Python brain
THE REWARD (Web3/Solana): Session completed successfully? Mint Solana tokens to your wallet Track earnings in real-time
Challenges we ran into
Ultrasonic Sensor Jitter: We initially struggled with "sensor noise," where the ultrasonic sensor would intermittently return 0 cm readings due to sound wave absorption or poor bounce angles (e.g., hitting a soft sweater). This caused the LCD to flicker between colors. We solved this by implementing a software-based state machine and signal filtering logic to ignore "impossible" readings.
I2C Bus Congestion: Spanning the I2C bus with constant color update commands caused lag and random characters on the LCD. We had to optimize the code to only update the backlight when a definitive state change occurred.
Real-time Synchronization: Getting a Python AI model (running at 30 FPS) to communicate smoothly with an Arduino (running at 9600 baud) without causing serial buffer overflows required careful timing and efficient command strings.
Accomplishments that we're proud of
The Hardware-Software Bridge: Successfully creating a low-latency feedback loop where a webcam's AI "vision" controls physical LCD in real-time.
Seamless Pose Detection: Fine-tuning the MediaPipe landmarks to accurately detect "slouching" versus "looking away" without being overly sensitive to natural head movements.
Proof of Focus Logic: Architecting a system that doesn't just track time, but verified time, ensuring that users can't "game the system" by leaving the frame or using a static image.
What we learned
Signal Processing Basics: How to implement simple software filters (like the "999 trick") to clean up noisy hardware data.
State Machine Architecture: The importance of using states (Idle, Focused, Away) to manage complex logic instead of long strings of nested if-else statements.
DePIN Concept Mapping: How to bridge physical human activity with blockchain incentives, effectively turning a standard desk setup into a decentralized infrastructure node.
What's next for Snitch It
Solana Mainnet Integration: Moving the reward system from a simulated trigger to an actual automated smart contract on the Solana mainnet.
Gamification: Developing a leaderboard system where users can "stake" their focus tokens to enter focus competitions with other builders.
Enhanced Enclosure: Designing a custom 3D-printed housing for a Logitech webcam and Grove components to make it a sleek, consumer-ready desktop device.
Log in or sign up for Devpost to join the conversation.