Packybara

View the Packybara product page →

Brief Overview

Packybara is a package tracking app designed for the Brazilian market, where tracking shipments from multiple carriers (Correios, AliExpress, Shopee, Anjun, J&T) typically means juggling different websites and apps. A pixel-art capybara mascot accompanies the user through each delivery journey with contextual animations and messages in Portuguese.

Available on iOS (TestFlight) and Android (Google Play closed testing).

Problem Statement

Brazilian online shoppers frequently order from multiple platforms — Correios for domestic, AliExpress and Shopee for international — each with their own tracking system. There’s no single, delightful app that consolidates all carriers with personality and charm.

Challenge & Scope

• Multi-carrier aggregation: Automatic carrier detection via regex patterns on tracking codes, with a unified API layer through RapidAPI/PacoteVicio
• Capybara personality system: CapyMessageMapper maps (status, latestEvent) to contextual sprite states, animations, and Portuguese messages via keyword matching
• OCR code scanning (iOS): Apple Vision framework scans photos of shipping labels, WhatsApp screenshots, and receipts to extract tracking codes automatically — no typing required
• Siri & App Intents (iOS): “Hey Siri, where’s my package?” — check status, add packages by voice, and trigger delivery automations via Shortcuts
• Spotlight indexing (iOS): Active packages are searchable system-wide via CoreSpotlight
• Pixel art rendering: 64x64 sprites rendered with nearest-neighbor filtering and optional 4-frame animation
• Push notifications: Firebase Cloud Messaging for real-time tracking updates via Supabase edge functions
• Clipboard intelligence: Detects tracking codes on the clipboard and suggests adding them automatically

My Role & Process

Solo developer. Designed the pixel art sprites, built the full Android app with Jetpack Compose + Material3 and the iOS app with SwiftUI + CoreData. Implemented carrier detection, OCR scanning via Vision framework, Siri integration via App Intents, tracking normalization layer, and CI/CD pipelines for both App Store and Play Store.

Key Technical Decisions

iOS

• SwiftUI + CoreData — MVVM with @Observable view models (iOS 17+ Observation framework)
• Apple Vision for OCR — .accurate recognition level with language correction disabled to preserve tracking code fidelity
• App Intents framework — wrapper entities (PackageAppEntity, CarrierAppEntity) keep domain models clean from framework concerns
• Cache-first Siri responses — intents read CoreData directly for sub-second answers, avoiding network calls

Android

• Ktor over Retrofit — lightweight, Kotlin-native HTTP client that plays well with coroutines
• Room for local persistence — offline-first with PackageEntity and TrackingEventEntity
• Hilt for DI — standard Android DI, keeps the single-module architecture clean

Shared

• PressStart2P font — reinforces the pixel-art identity across both platforms
• Supabase for backend — auth, postgrest, and edge functions without managing infrastructure

Outcome

iOS app on TestFlight with full Siri integration, OCR scanning, and Spotlight indexing. Android app in closed testing on Google Play. Both apps support all major Brazilian carriers and feature a fully animated capybara companion. Positioned for public launch on both platforms ahead of WWDC 2026.