Skip to content

sephorah/mpc-hacks

Repository files navigation

CareWorkspace

Zero-wait-time async virtual care — built at MPC Hacks 2026 (Polytechnique Montréal)

CareWorkspace rethinks where virtual care actually gets stuck. The clinical encounter is fast; the friction lives in everything before the clinician arrives. Instead of routing everyone into the same slow live-video queue, CareWorkspace asks a different question: can this case close asynchronously right now, and if not, what's missing to close it?


How it works

A patient files a short structured intake — service type, a set of yes/no red-flag safety questions, and document readiness. Deterministic rules (not the AI) read that intake and assign one of three lanes:

Lane Condition What happens
needs-sync Any red flag triggered Routed to a live visit
async-pending Missing document Parked until the patient supplies it
async-ready All clear Queued for async clinician review

Google Gemini summarises the patient's intake into a short decision brief. If the model is unavailable, a deterministic fallback surfaces the raw intake so the case can still move forward. The clinician reviews like-with-like (cases are grouped by type on their dashboard), reads the brief, enters their name, and attests to close — in under a minute. The patient is notified once the case is reviewed.

The model makes zero clinical or routing decisions. Every safety gate is deterministic. The AI's only job is to reduce reading time, a human makes all final decisions.


Tech stack

  • Framework: Next.js 16 + React 19, TypeScript throughout
  • Styling: Tailwind CSS v4
  • Server: Custom tsx-based server
  • AI: Google Gemini 2.5 Flash via @google/genai
  • State: In-memory
  • Linting / formatting: Biome
  • Deployment target: Vercel

Getting Started

First, run the development server:

npm run dev
# or
yarn dev
# or
pnpm dev
# or
bun dev

Open http://localhost:3000 with your browser to see the result.

You can start editing the page by modifying app/page.tsx. The page auto-updates as you edit the file.

This project uses next/font to automatically optimize and load Geist, a new font family for Vercel.

Learn More

To learn more about Next.js, take a look at the following resources:

You can check out the Next.js GitHub repository - your feedback and contributions are welcome!

Deploy on Vercel

The easiest way to deploy your Next.js app is to use the Vercel Platform from the creators of Next.js.

Check out our Next.js deployment documentation for more details.


Design principles

Safety is the first gate, not a bolt-on. Any red flag in the intake immediately routes to a live visit — no model involvement, no override path.

The model assists; rules and humans decide. Gemini writes a summary brief. The clinician reads it and attests. Routing, classification, and closure are all deterministic or human-gated.

Resolution over routing. The system is designed to close cases, not just queue them. The async lanes exist to eliminate wait time for cases that don't need a live slot, not to defer them indefinitely.


What's next

  • Tighter, more specialised intake summaries to improve brief quality and reduce error margin
  • Direct patient file attachment to eliminate document round-trips
  • Rules-defined cohorting — reviewing a pre-screened homogeneous group and attesting once — as the path from hundreds of cases a day toward thousands

Built at MPC Hacks 2026 for Dialogue's Zero Wait Time Virtual Care challenge.

This is a Next.js project bootstrapped with create-next-app.

About

MPC Hacks 2026 — invited to present the project at Dialogue's internal hackathon.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors