Mealcraft

Inspiration

We've all been there - staring into a fridge full of random ingredients with no idea what to make. Meanwhile, recipe apps expect you to already know what you want to cook.

We wanted to flip the script.

Mealcraft was born from a simple question: What if your phone could look at what you have and tell you what to make?

We were also frustrated by the recipe discovery experience.

You find an amazing dish on TikTok or Instagram, but saving it means losing it in a sea of bookmarks.

We wanted one place where all your recipes - from grandma's cookbook, that viral reel, or a random food blog - could live together.

What does it do?

Mealcraft is your AI-powered kitchen companion that:

Scans your ingredients - Point your camera at your fridge or pantry, and AI identifies what you have.
Generates personalized recipes - Based on your ingredients, dietary preferences, and cooking skill level.
Extracts recipes from anywhere - Paste a link from TikTok, Instagram, YouTube, or any food blog, and we'll pull out the recipe automatically.
Digitizes physical cookbooks - Snap a photo of a cookbook page, and we'll extract the recipe.
Plan your meals - Organize what you're cooking for breakfast, lunch, and dinner.
Builds grocery lists - Automatically generates shopping lists from your planned recipes.

How we built it

Mobile App: React Native + Expo
Navigation: Expo Router with tab-based navigation
State Management: Zustand + TanStack Query
Backend: Node.js API routes
Database: PostgreSQL (Neon)
AI/ML: Google Gemini for vision & text generation
Payments: RevenueCat for subscriptions
Styling: Custom theming with dark mode support

Challenges we ran into

Recipe Extraction from Social Media

Social platforms don't make it easy to extract content. Instagram and TikTok have strict APIs, so we had to get creative with web scraping, caption extraction, and AI interpretation.

Each platform required its own extraction strategy.

Ingredient Recognition Accuracy

Getting AI to reliably identify groceries from photos was tricky. A tomato in bad lighting could be mistaken for an apple.

We fine-tuned our prompts extensively and added user confirmation steps.

Offline-First UX with Device IDs

Without traditional user accounts, we needed to persist data across sessions. We implemented device-based identification while planning for future account sync capabilities.

Cooking Mode Experience

Designing a hands-free cooking interface that's usable with messy hands required rethinking typical mobile UX patterns.

Large tap targets, voice considerations, and preventing screen timeout were all factors.

Accomplishments that we're proud of

Universal recipe import - One app that handles links, photos, and manual entry
Smart grocery aggregation - Combines ingredients across multiple planned recipes
Delightful onboarding - Personalized experience based on cooking skill and dietary needs
Beautiful dark mode - Full theme support throughout the app
Sub-second ingredient scanning - Real-time camera feedback

What we learned

Mobile camera UX matters - Users need clear feedback about what they're scanning
Recipe data is messy - Every food blog formats differently; normalization is essential
Offline-first is hard - Syncing local and remote state requires careful architecture