PanMe

Inspiration

Two deeply personal experiences inspired me to build PanMe. First, food apps have been my creative playground for years. During COVID, while everyone was baking sourdough, I was frustrated planning my bread bakes in Excel spreadsheets. So I did what any developer would do, I learned new technology and built my first app: a bread baking planner. What started as a solution for myself turned into something people genuinely loved. It's been in the top 10 of the food and drinks category on and off for four years now.

Then came my second food app, MealMe, a meal planner my wife and I wanted for ourselves. Friends started using it so much that I knew I had to release it properly. These two apps taught me something crucial: the best solutions come from solving real problems you experience yourself.

So when Eitan's brief came up for the hackathon, I jumped on it immediately. I already understood the food app space deeply. I knew the pain points.

What it does

PanMe is a cookbook app and smart shopping list combined into one seamless experience. It removes friction at every step of the cooking journey. From discovering recipes to shopping to actually cooking hands-free.

Key features:

Smart Recipe Importer: Import recipes from blogs or YouTube videos automatically. No more copying and pasting ingredients or trying to keep a video open while you cook.

Cooking Term Tooltips: Built-in explanations for cooking terminology. When you see "julienne" or "fold" in a recipe, just tap to understand exactly what it means. This is the feature I wish I'd had when I started cooking.

Hands-Free Cook Mode: Voice-controlled navigation through recipe steps. Just say "next" or "back" and the app moves through pages, no need to touch your phone with greasy fingers.

Smart Shopping List with Tips: Ingredients automatically populate your shopping list with helpful shopping tips. You can rearrange categories to match your actual store layout, so ingredients appear in the order you'll find them while shopping. No more backtracking through aisles.

How I built it

I built PanMe as a solo developer using my go-to iOS development stack:

• Swift and SwiftUI for the native iOS interface
• SwiftData for local data persistence
• RevenueCat for subscription and paywall management
• OpenAI API for intelligent recipe parsing and cooking term explanations

The architecture prioritizes a clean, intuitive user experience. The recipe importer uses the OpenAI API to intelligently parse recipe content from various sources, extracting ingredients, instructions, and timing information. The smart shopping list aggregates ingredients across multiple recipes, intelligently combining duplicate items and summing quantities.

The hands-free Cook Mode was designed with real cooking scenarios in mind, when your hands are covered in flour or handling raw chicken, voice control becomes essential, not just convenient.

Challenges I ran into

The biggest challenge was actually a pivot from my original vision. I initially wanted to use the iPhone's proximity sensor for hands-free navigation. You'd simply wave your hand near the phone to move to the next step. Elegant, right?

Wrong. The screen would turn off every single time the proximity sensor activated, and there's no way to override that behavior in iOS. After hours of trying different approaches, I had to admit defeat and rethink the entire interaction model.

That's when voice mode was born. And honestly? It turned out to be much better than my original idea. Voice control is more reliable, works from any distance, and feels more natural when you're actually cooking.

The second major challenge was ingredient mapping and aggregation. Recipes describe ingredients in wildly different ways, "2 tomatoes," "3 medium tomatoes," "tomatoes, diced," "1 can diced tomatoes." Getting the shopping list to intelligently recognize these as the same ingredient (or different ingredients) and properly sum quantities was surprisingly complex. But solving this was essential, nobody wants a shopping list with five separate entries for tomatoes.

Accomplishments that I'm proud of

I'm genuinely proud of the voice mode. What started as a fallback solution became the app's signature feature. It makes Cook Mode feel effortless and actually useful in real cooking scenarios.

The Cook Mode interface itself is something I'm really happy with, it helps you focus on exactly what you need to do in the moment, without distraction. Clean, simple, effective.

But the feature I'm most proud of might be those cooking term tooltips. It's such a small thing, but it's the feature I wish I'd had when I was learning to cook. Not everyone knows what "fold" or "bloom" means, and having instant explanations removes that moment of hesitation and uncertainty.

What I learned

Recipes can take many shapes and forms, and it's genuinely hard to account for all cases. Every food blogger has their own style, every video has different structure, every cookbook formats things differently. Building a system that handles this variety is more art than science.

What I built for this hackathon is a great starting point, it handles the most common cases well, but there's so much room to improve. And that's exciting. I learned that sometimes the best approach is to ship something good and iterate based on real usage, rather than trying to solve every edge case upfront.

I also learned (again) that the best features often come from constraints and failures. Voice mode only exists because proximity sensing didn't work. That limitation forced a better solution.

What's next for PanMe

My immediate priority is improving recipe extraction to make it more robust and reliable. I want to expand import support to TikTok, Instagram, and other platforms where people discover recipes today. The current system works well, but making it handle even more sources and edge cases will make PanMe truly versatile.

I'd also love to partner with someone to fill the app with curated cookbook content. Right now it's great for importing recipes you find, but I want PanMe to also be a destination for discovering new recipes, trusted collections you can explore and cook from.

Beyond features, I'm excited to see how real cooks use PanMe and what friction points I haven't anticipated yet. The best product insights always come from actual usage, not assumptions.

Built With

• swift
• swiftdata
• swiftsoup
• swiftui
• thecomposablearchitecture