CARe - the AI-driven hospital navigator

CARe is an AI-driven Hospital Navigator. We believe that the lack of doctors is a big problem in Malaysia and it leads to problems, one of which is, the long waiting-time. To counter this problem, this app takes in factors such as a hospital’s estimated waiting time (based on the number of people already at the hospital) and the distance between patients and hospitals to recommend the hospital that provides the shortest time to diagnosis.
The app also collects a short, AI‑guided intake on the way and produces a clean PDF summary that can be shown to doctors or shared with clinicians (QR/link).

Judging criteria (summary)

• Innovation: Combines travel ETA with live/estimated wait; on-the-way AI intake; instant PDF/QR handoff.
• Feasibility: Built with proven components (FastAPI/Next.js, Google Maps APIs); mock mode enables immediate trials.
• Technical merit: Async calls to Places/Routes, optional Vision, in-memory + MySQL cache, typed models and clean API boundaries.
• User benefit: Reduces total time-to-care and improves first contact quality via structured intake.
• Deployability: Frontend deployable on Vercel; backend runs on any VM/container with simple env vars; DB optional.

Technologies and resources

• Frameworks: FastAPI, Next.js 14 (App Router), React 18, TypeScript.
• Libraries: SWR, Tailwind, framer-motion, lucide-react, pdf-lib, sqlalchemy, requests, python-dotenv.
• Databases: MySQL (optional cache via SQLAlchemy).
• APIs: Google Places API, Google Routes (Distance Matrix computeRouteMatrix).
• AI models/services: Groq (chat/summary, default model configurable), OpenAI Vision (optional) for headcount.
• Prompting: Simple task prompts for Groq chat and a JSON-only counting prompt for Vision
• Datasets: Demo public/mock/hospitals.json for offline/mock mode.