DisPill
A Secure, Accessible, AI-Powered Medication Dispensing System
🚀 Introduction
DisPill is a smart medication dispensing platform designed to make medication management safer, more accessible, and more reliable for everyone.
Medication non-adherence is a serious global problem—missed doses, accidental overdoses, and unauthorized access to prescription drugs put lives at risk. DisPill tackles these challenges with a seamless blend of hardware, AI, and thoughtful design.
Inspiration
Medication non-adherence is a serious global problem—missed doses, accidental overdoses, and unauthorized access to prescription drugs put lives at risk. We decided to build a system that securely solves these problems without being intrusive.
What it does
Our system:
Reminds users when to take medications, reducing missed or late doses
Physically dispenses pills, minimizing effort for users with disabilities or limited mobility
Secures sensitive medications using biometric authentication through Med Lock, preventing unauthorized access to addictive or dangerous drugs
Provides users with a personalized medication assistant we used Gemini and 11Labs to help provide users with an AI assistant that can help them with their medications via speech
Tracks inventory and refill dates, ensuring users never run out unexpectedly
Store medication history, gives users the chance to look back at their medication history, and DisPill leverages AI to summarize the history for users.
To simplify onboarding, we use Google Gemini to automatically extract structured data from prescription images or PDFs. This data powers a personalized AI medication assistant, enhanced with ElevenLabs voice synthesis, allowing users to interact with DisPill in a natural, conversational way.
DisPill isn’t just a reminder app—it’s a full medication safety ecosystem built with empathy, security, and real-world usability at its core.
How we built it
- As students we understand the usefulness of VIBE CODING but we stayed off that route as much as we can so that we can have a learning experience out of the event.
- The frontend and backend was done with React(NextJS, Typescript).
- The database and backend services were built with Firebase.
- We used Gemini and 11Labs as AI integrations into the application.
- Claude AI was used to implement some functions and dummy features/interfaces.
- We built the Pill Dispenser with servos, arduino, RedBull cans, cardboard, tapes, and used M&Ms as our pills.
🛠 Built With
React.js
Next.js
TypeScript
Firebase
Google Gemini API (Prescription data extraction)
ElevenLabs API (Voice-based AI assistant)
Claude AI
Arduino
Servo Motors
(Yes) Red Bull cans, tape, and a lot of love & care ❤️
Challenges
- The hardware was one major problem for us, we needed a 3D printer and soldering station but since we couldn't access any, we decided to use RedBull cans and a simple arduino kit to achieve our goal.
- With the software, we needed to run some jobs to make sure users receive their reminder emails and notifications on time. For the frequency we needed to run the jobs, we couldn't find a good and easy to use option, so we ended up running a function periodically on our own local server. It "works" 😊.
- We wanted to issue remote commands to the hardware for when to start giving out pills, but because we used an arduino and not a raspberry PI, the software capability of our hardware was limited.
- Our Hardware successfully dispenses drugs and we were happy but by enjoying it too much we ended up weakening some parts so an inaccurate number of pills can be dispensed.
We are proud of
Our hardware: it was a big headache at first when we built what we wanted with a Raspberry PI. But we had so much fun working with the boards and cans and playing around with the degrees we should turn the servo, we came up with a bunch of random design ideas, and at the end it "worked".
Our UI: we put a lot of work into making our UI User-Friendly and clean but playful enough to keep users engaged (Which in turn means they take their medications). We put work into some creative animations and UX decisions and we are proud of that.
What we learned
As a team, we learned teamwork. We learned to help each other. We all came together with very different strengths, one of us (maybe me ~the typist) was invited into the group even after stating he has low technical abilities. We've learned to help each other grow, and individually, we've learned a lot from each other.
What's next
- We are going to build a better pill dispenser and make the pill dispenser very tough. If it is not strong enough, a software lock can't keep a loved one safe from it's content.
- We are going to design variants of the hardware to suit specific groups of individuals to make it more accessible.
- We are going to make the software very very secure. Health data is very personal and sensitive.
Team Members
- Vineel Kundula
- Jaeyoung Choi
- Benjamin Akusaki
- Micheal Owusu
Built With
- arduino
- claude
- elevenlabsapi
- firebase
- geminiapi
- next.js
- react.js
- servomotor
- typescript


Log in or sign up for Devpost to join the conversation.