Inspiration
We all have a version of ourselves we aspire to be: the "Kitchen Pro" who effortlessly whips up the incredible meals we see online. We feed this dream by saving recipes—thousands of screenshots, bookmarks, and videos. But in reality, that digital hoard becomes a graveyard. The sheer friction of finding that one recipe across five different places paralyzes us, and we end up ordering takeout again. I realized the problem wasn't a lack of desire or recipes; it was a lack of executive function. We didn't need another place to find food; we needed a system to help us execute the food we already found. Eitan Bernath’s brief confirmed this: the goal was to take users from "passively saving" to "actively cooking." Aipron was born to bridge the gap between the dream of cooking and the reality of dinner.
What it does
Aipron is the Operating System for your Kitchen. It acts as a centralized HQ that aggregates your scattered recipes into one actionable database.
- Universal Import: Whether it’s a photo of Grandma’s handwritten note, a messy PDF, or a YouTube video, Aipron scans and structures it into a clean, interactive recipe card.
- Smart Grocery List: This is the bridge between planning and doing. You can add ingredients from any recipe directly to your shopping list with a single tap. The app intelligently aggregates items (e.g., combining "2 onions" from one recipe and "1 onion" from another into "3 Onions"), so you know exactly what to buy without the mental math.
- Contextual AI Chef: Stuck with ingredients? You can chat with the app offline. Tell it what you have in the fridge, and it suggests meals from your own library.
- Cooking Mode: A distraction-free, step-by-step interface that guides you through the process, preventing the "messy hands on screen" struggle.
How we built it
We built Aipron with a "Local-First" philosophy using Expo React Native. Speed and privacy were non-negotiable for a tool used in a busy kitchen.
- The Brain: We used Google Gemini 2.5 Flash for the heavy lifting—parsing messy PDFs and unstructured screenshots into structured JSON data.
- The Privacy: For the conversational AI and search, we implemented Executorch running Llama 3.2 1B directly on the device. This ensures users can chat with their recipe database entirely offline, keeping their data private.
- The Business: We integrated RevenueCat to manage entitlements for our Pro tier, ensuring a seamless subscription experience.
Challenges we ran into
The biggest hurdle was teaching the app to understand "messy" human intent. Recipes come in non-standard formats—cursive handwriting, chaotic blog layouts, or spoken words in a video. Fine-tuning the prompt engineering to reliably extract structured data (separating ingredients from instructions) was difficult but necessary. If the data wasn't clean, the Smart Grocery List wouldn't work, and the friction would return. We had to iterate extensively on our parsing logic to ensure that "1 cup of diced carrots" was understood as an ingredient to be bought, not just text to be read.
Accomplishments that we're proud of
We are incredibly proud of the "Zero Friction" Onboarding. We resisted the urge to add login screens or cloud syncs. The app loads instantly, works offline, and requires no account creation. We are also proud of the On-Device AI implementation. Getting Llama 3.2 running locally on a mobile device to query a custom JSON database was a significant technical achievement that proves you don't need a massive server bill to build intelligent apps.
What we learned
We learned that technology is only useful if it changes behavior. By removing the barriers of entry—instant import, smart aggregation, offline capability—we found that users actually stopped hoarding and started cooking. We learned that the "Dream Cook" isn't a fantasy; they just needed the right tool to wake them up.
What's next for Aipron: Recipe App
The next phase is about transforming Aipron from a tool into a companion. • Enter "Aipron": We are implementing our mascot, "Aipron" (The AI Apron). This isn't just a logo; it’s a "smart garment" personality that reacts to your activity. It brings an emotional layer to the app, cheering you on when you cook and nudging you when you’ve been ordering too much takeout. • Empathetic AI: We are refining the on-device AI to deeply understand dietary nuances. You won’t just ask for a recipe; you’ll be able to say, "Make this dairy-free," and the AI will intelligently adapt your saved recipes to fit your needs. • Gamified Community: We are introducing Cooking Challenges to turn meal prep into a shared victory. Users can join community streaks (e.g., "7 Days of Home Cooking") to build habits together. • The Indie Chef Spotlight: We are launching a feature to collaborate with independent culinary creators. This dedicated section will spotlight their unique recipes, bringing high-quality, creator-driven content directly into your personal HQ.
Built With
- android
- executorch
- expo.io
- gemini
- ios
- llama
- react-native
- revenuecat
- sqlite
- typescript
Log in or sign up for Devpost to join the conversation.