CareCompanion
A private health dashboard for family caregivers: medication tracking, vitals, symptoms, appointments, and AI-powered medical report parsing. 100% local data — not a single byte of health information leaves the device.
- Next.js
- TypeScript
- Groq AI
- Llama-3.3-70b
- PWA
- Tailwind CSS
- Healthcare
01 / Result
- 1st
- Place
- 11
- Projects
- 25–30
- Participants
- 0
- Bytes uploaded
First place at a health-tech hackathon, out of 11 submitted projects.
02 / The gap
Problem
Family caregivers juggle medications, vitals, symptoms, and appointments across notebooks and group chats, with no centralised tool built for their needs — and no guarantee of privacy if one existed.
Approach
A progressive web app with an AI health assistant and PDF report parsing, built on a strict localStorage-only architecture so sensitive health data never leaves the caregiver's own device.
03 / Features
- Medication tracking — schedules and adherence in one place, so nothing gets double-dosed or missed.
- Vitals & symptoms — log readings over time and see the trend, not just today's number.
- Appointments — a single timeline across every clinician involved in care.
- AI report parsing — drop in a PDF medical report and get a plain-language summary, powered by Groq running Llama-3.3-70b.
- Installable PWA — works like a native app on phone or desktop, including offline.
04 / On privacy
Health data is the most sensitive category of personal information there is, and caregivers are handling someone else's. The architecture makes the guarantee structural rather than a promise: there is no backend to leak, because everything lives in the browser's own storage on the device you installed it on.