Trotty

Problem

Dog owners want a simple way to track walks — distance, route, bathroom and sniff markers — across their phone and watch. But a walk route is a continuous record of where you live and go every single day. Most tracking apps ship that to a server in plaintext. I wanted live tracking without handing over a map of someone's daily life.

My Role

Architecture, native iOS + watchOS development, on-device cryptography, the GPS tracking engine, and the auth backend.

Process

1. 1

   Built the tracking engine on CoreLocation with GPS smoothing, dwell detection, and sniff/altitude markers, rendering live routes on MapKit; persistence and the data model run on SwiftData.

2. 2

   Encrypted every location point on-device with CryptoKit (AES-GCM), generating and storing keys in the Keychain — the server never sees a plaintext track. Wrote dedicated tests for the crypto engine and track-point encryption.

3. 3

   Built a watchOS companion that records walks independently and syncs to the phone over WatchConnectivity, plus Live Activities (ActivityKit) and Home Screen widgets (WidgetKit) for at-a-glance walk status.

4. 4

   Backed Sign in with Apple with a stateless Cloudflare Worker that mints short-lived ES256-signed JWTs from the team's .p8 key to call Apple's token and revoke endpoints — no database, no logging, no persisted state.

Outcome

A V1 in development spanning iPhone, Apple Watch, widgets, and Live Activities — with end-to-end on-device encryption of all location data and a zero-state auth backend. The architecture keeps the most sensitive data (where you walk every day) on the device by default.

Stack

What I Learned