PetPilot

Winner, AI Valley Dogathon 2026. Built in 5 hours by a team of 4.

Shelter software has a field for everything that fits in a form and nothing for what staff actually notice. A volunteer sees a dog pee purple, or notice it was lethargic on a short walk, and there is nowhere structured to put that. It ends up in a group text, or nowhere. Even when it does get written down, there is no clear rule for who should hear about it: sometimes the vet team, sometimes behavior, sometimes facilities, and the shelter staff we talked to said there is no one clear-cut way to decide.

PetPilot is a shelter management front end plus a remarks router: type the observation in plain English, and it gets classified, routed to the right teams, and actually sent.

How it works

browser (the Foster Board page)
   |  open a dog's Remarks, type an observation, hit "Classify & route"
   v
remarks-router backend  (localhost:3001)
   |  1. Claude classifies the text into one of six labels
   |  2. a routing table maps the label to the responsible teams
   |  3. Arcade sends the alert to Slack, or logs it in log mode
   v
a new row appears in the dog's Remarks, tagged and routed

Example: "peed purple this morning, seemed fine otherwise" classifies as medical_urgent, which routes to the teams responsible for urgent medical calls.

Routing is a lookup table from label to teams, not a model decision. That keeps the behavior inspectable: you can read exactly why an alert went where it went, which matters more than sophistication when a vet is the one being paged.

Arcade handles the authenticated send, so the project does not ship its own OAuth flow and per-user tokens stay out of this codebase. That is deliberate: a shelter alert router has to reach people across a volunteer base that turns over constantly, so delegated auth is the right primitive rather than an afterthought.

Live version

Deployed on Replit at https://dog-os-dogathon-1.replit.app/.

This repo is the source of record.

Running it locally

You need Node.js and Python. Two terminals, both stay open.

1. Backend

cd remarks-router
cp .env.example .env      # Windows: copy .env.example .env
npm install               # one time
npm start                 # prints "remarks router listening on :3001"

Fill in .env before starting:

Variable	Needed for
ANTHROPIC_API_KEY	Required. Without it, classification fails. Get one at console.anthropic.com
ARCADE_API_KEY / ARCADE_USER_ID	Only for real Slack sending
SEND_MODE	log classifies and logs the alert without sending. slack sends for real, once Arcade is set up and you have joined the channels

.env is git-ignored. Never commit it. Each teammate makes their own.

Windows: if npm fails with "running scripts is disabled," use Command Prompt instead of PowerShell, or run once: Set-ExecutionPolicy -Scope CurrentUser -ExecutionPolicy RemoteSigned

2. Frontend (second terminal)

python -m http.server 8000

From the repo root, then open http://localhost:8000/Foster%20Board.dc.html.

The page must be served, not opened as a file. Double-clicking the HTML leaves every {{ value }} unresolved because the runtime cannot boot.

3. Try it

Open a dog's Remarks, type an observation, hit Classify & route. A new row appears with a label chip (red for medical_urgent) and the teams it routed to.

If you see "Couldn't reach the router," the backend is not running or errored. Check the backend terminal; a missing ANTHROPIC_API_KEY shows up there as an authentication error.

Where things live

Path	What
Foster Board.dc.html	The whole frontend app: foster roster, medical calendar, intake and outcomes, and the Remarks drawer. Single-file Design Component app, React via CDN
support.js	The Design Component runtime the page loads
_ds/	Design system: styles and component bundle
remarks-router/	The classify, route, and send backend (Node/Express). See its own README for endpoints and internals

The frontend calls http://localhost:3001, set once as REMARKS_API near the bottom of Foster Board.dc.html. Change that one line to point elsewhere.

Honest limits

Built in one 5-hour session, so:

• All shelter data is synthetic. No real PetPoint integration; this is designed as a layer that would sit alongside a shelter's system of record, not replace it.
• Routing is a fixed table, not learned from a shelter's own history of who got alerted for what. That is the obvious next step and the reason to want real data.
• Six labels is a starting taxonomy, not a validated one.

What this would unlock

The router has to touch every system a shelter already runs, which makes it a reasonable wedge for the things the shelter staff asked for and we could not build in five hours: syncing to electronic health records so an adopted animal arriving at a hospital does not need manual re-entry, keeping Salesforce and the shelter system consistent instead of diverging, inter-shelter transfer requests with tracking, and quarantine history that is currently hard to retrieve at all.

Team

• Pranav Balachander, pruhnav
• Kiran Oguri, kiran172
• Arshia Bhattacharyya, arshiabyya
• Trisha Moorkoth, tprofessional