PackPal

PackPal is your smart travel assistant that predicts what you’ll need for any trip by analyzing your destination, weather, and plans, creating personalized, effortless packing lists in seconds.

Comment

Inspiration

Packing for trips is always stressful, with overpacking, forgetting essentials, or not knowing what to bring based on the weather or activities. I wanted to create an AI-powered assistant that removes the guesswork, making travel preparation effortless, smart, and quick

What it does

PackPal automatically generates a personalized packing list for any trip. You simply enter your destination, dates, and trip type (e.g., vacation, business, camping). PackPal then uses: Real-time weather data Activity detection from flight or trip details Local context and smart AI suggestions to build a packing list that fits you and your trip, no overpacking, no forgetting.

How we built it

We built PackPal using SwiftUI for the iOS app, integrating several intelligent modules: OpenWeather One Call API for dynamic climate analysis. Core ML + MobileBERT for on-device AI embedding and text understanding. Custom AI generation logic trained to infer essentials from destination, weather, and duration. A modular rule-based fallback system that ensures offline performance. We converted and integrated Core ML models for Apple devices to ensure zero latency and full privacy, no cloud required!

Challenges we ran into

Model conversion issues: Converting transformer models like DistilBERT to Core ML introduced scriptability errors. Performance trade-offs: Balancing a lightweight inference with accurate results. Weather + activity mapping: Creating a logic that adapts the suggestions for varying conditions (e.g., humid + beach vs. cold + business trip). Offline AI constraints: Ensuring the app remains useful without internet access.

Accomplishments that we're proud of

Achieved a fully functional local AI system, no internet or expensive APIs needed. Built a full packing generator that factors in real-world travel data. Created a smooth, modern SwiftUI interface (with iOS 26 Liquid Glass) with live feedback and smart suggestions. Integrated Core ML and custom AI workflows in a user-friendly app.

What we learned

How to integrate machine learning pipelines into iOS apps effectively. The importance of AI fallback systems. How to convert and deploy transformer models on-device. That simplicity and user experience often matter more than complex cloud systems. How to adopt Apple's new Liquid Glass

What's next for PackPal

Add trip-sharing and collaborative packing (for group travel). Expand AI personalization: learn from user behaviour over time. Integrate flight and itinerary parsing (auto-detect destinations and trip lengths). Launch on TestFlight, then on the App Store for public use. Experiment with vision-based packing validation using the camera (confirming packed items visually).

Built With

Share this project:

Updates