NANI: Your AI Medicine Companion

An AI-powered medication adherence system built on Raspberry Pi, Infrared sensors, and real-time health alerts. "Nani" means grandmother in Hindi - a symbol of warmth, care, and gentle guidance.

Comment

Inspiration

We are inspired by our parents and grandparents - the people who taught us care, patience, and love. As they grow older and manage multiple medications, simple routines become overwhelming. Missed doses can lead to anxiety, hospital visits, or worse. We thus wanted to create something that felt as caring and reliable as a family member, a system that blends AI and IoT to make medicine time gentle, not stressful. That’s how NANI came to life - a compassionate, tech-driven companion for everyday health.

What it does

NANI helps elderly users stay on top of their medication schedules while keeping caregivers connected.

  • Smart sensing: Detects when a pill compartment is opened using IR break-beam sensors.
  • Real-time tracking: Sends timestamps to the cloud backend, creating a reliable medication log.
  • Smart reminders: If a dose is missed, NANI issues voice or app-based alerts.
  • Caregiver notifications: The iOS app updates family members or doctors instantly.
  • Voice AI: Enables conversational check-ins like “Did I take my pill?” or “Remind me in an hour.”

How we built it

  • Hardware: Raspberry Pi 3 Model B connected to an IR break-beam sensor mounted on a standard pill organizer.
  • Firmware: Python scripts detect beam breaks and send POST requests to our backend API.
  • Backend: Node.js + Express server managing state, timing logic, and alerts.
  • Cloud Integration: Google Apps Script logs and syncs real-time events for visibility.
  • Frontend: Swift iOS app for caregivers to track adherence and send reminders.
  • Voice Assistant: Connected to OpenAI API for auditory feedback and gentle reminders.

Challenges we ran into

  • Calibrating the IR sensor to accurately detect pillbox openings without false positives.
  • Managing power and GPIO pin limitations on the Raspberry Pi.
  • Debugging real-time event sync between local Pi sensors and the cloud backend.
  • Network instability during testing - ensuring data still logged reliably offline.
  • Integrating AI APIs into XCode while using Swift.
  • Building a cohesive end-to-end system (hardware, backend, and mobile app) in just under 36 hours.

Accomplishments that we're proud of

  • Built a working IoT prototype that detects pillbox usage and triggers live cloud updates.
  • Achieved a real-time, hardware-to-mobile pipeline using lightweight APIs.
  • Created a caregiver notification system that adds emotional assurance to technical reliability.
  • Integrated voice AI for a friendly, accessible interface for elderly users.

What we learned

  • How to use Raspberry Pi GPIO for precise IR sensor input and event handling.
  • Designing fault-tolerant data flows between hardware and cloud systems.
  • The value of balancing tech efficiency with user empathy in healthcare design.
  • API Integration, Testing, and Improvement
  • How to rapidly iterate across hardware, software, and UX under time constraints.

What's next for NANI: Your AI Medicine Companion

  • Pill recognition using computer vision to track dosage automatically.
  • Predictive adherence analytics to alert caregivers before a missed dose occurs.
  • Expanded hardware support for additional sensors (weight, touch, motion).
  • Vetted conversational health AI for mood check-ins and medication education.
  • Secure medical cloud with patient privacy and HIPAA compliance.

Contact Info

NANI’s mission: to make compassionate technology a daily part of healthcare, helping people live independently, safely, and with dignity.

Thank you for reading! We're excited to share our work with you:)

Built With

Share this project:

Updates