HackOMania 2026 - PAB Triage MVP

AI-assisted triage dashboard for Personal Alert Button (PAB) alerts, built for HackOMania 2026.

This project is a monorepo MVP that helps operators monitor incoming alert cases, review AI-assisted triage outputs, and make faster response decisions using resident context, call history, and audio-derived signals.

---

Team

Chia Yuan Jun

• GitHub: @cyuanjun
• LinkedIn: Chia Yuan Jun

Lee Jinseo (Jason)

• GitHub: @JSL124
• LinkedIn: Jinseo Lee

Su Yi Maung (Monica)

• GitHub: @Monica-Cheng
• LinkedIn: Monica Cheng

---

Overview

When a PAB alert is triggered, operators often need to make decisions quickly with incomplete information. This project aims to support that workflow by combining:

• Resident context from structured records
• Audio analysis from prerecorded or uploaded clips
• Language / dialect handling
• Rule-based triage fusion
• Operator review and final action tracking

The current implementation is designed as a modular MVP, with clean separation between frontend, backend, and persistence so it can evolve into a production-ready system later.

---

Key Features

• Operator dashboard for viewing and managing alert cases
• FastAPI backend for case handling and API services
• Next.js frontend for a clean web-based interface
• Resident context lookup from CSV-backed repositories
• Case persistence using JSON files
• Audio upload pipeline scaffold
• Language detection, transcription, and translation hooks
• Derived medical / history flags
• Fusion-based triage engine
• Case workflow state transitions
• Future-ready repository abstraction for Supabase or database migration

---

Repository Structure

.
├── backend/     # FastAPI service, repositories, tests, local data storage
├── frontend/    # Next.js + TypeScript + Tailwind dashboard
├── docs/        # Architecture notes
└── EXTRA/       # Extra materials / supporting files

Structure

• frontend/: Next.js + TypeScript + Tailwind dashboard
• backend/: FastAPI service with CSV repositories and JSON case store
• docs/: architecture notes

Quickstart

1) Backend

cd backend
python -m venv .venv
.venv\Scripts\Activate.ps1
pip install -r requirements.txt
uvicorn app.main:app --reload --port 8000

Optional (enables real transcript + language detection + translation):

$env:OPENAI_API_KEY = "sk-..."
# Optional model overrides
$env:OPENAI_WHISPER_MODEL = "whisper-1"
$env:OPENAI_TRANSLATION_MODEL = "gpt-4o-mini"
$env:OPENAI_SUMMARY_MODEL = "gpt-4o-mini"

2) Frontend

cd frontend
npm install
npm run dev

Frontend now calls a built-in same-origin proxy route by default (/api/proxy/*), which forwards to backend origin.

Optional backend origin override for the frontend proxy route:

$env:BACKEND_API_ORIGIN = "http://127.0.0.1:8000"

Optional direct API base override (skip proxy):

$env:NEXT_PUBLIC_API_BASE_URL = "http://localhost:8000/api/v1"

Tests

cd backend
pytest