Inspiration
Singapore is becoming a super-aged society. Seniors want to live independently, but bowel issues and undetected changes increase discomfort and risk. Caregivers need early, objective signals that respect dignity and privacy.
What it does
PooLabs is a discreet toilet-clip that privately detects bowel events and surfaces simple trend labels (no photos). Caregivers see only what matters: consistency trends and check-in prompts—nothing embarrassing, nothing stored.
How we built it
ESP32-CAM + one switch, and a battery. Is used to capture an image of the payload before being uploaded to our AI website to view, analyse and gain insights on the stool image.
Challenges we ran into
Balancing accuracy vs privacy, lighting/color drift in bathrooms, avoiding false positives (cleaning/turbulence), and keeping our public claims precise.
Accomplishments that we’re proud of
We shipped a calm, elder-friendly UX and an easy to setup hardware device.
What we learned
Great elder tech is invisible: fewer taps, fewer images, more meaning. Good heuristics + thoughtful hardware design often beat premature ML.
What’s next for PooLabs
On device computer vision, higher accuracy insights, and better hardware design.
Built With
- esp32-cam
- next.js
Log in or sign up for Devpost to join the conversation.