Sous: Turning Recipe Inspiration into Real Meals

Built for the RevenueCat Shipyard: Creator Contest (Eitan Bernath Brief #1)
Submission Period: January 15 – February 12, 2026

Hackathon Development

Sous was built specifically for this hackathon in response to Eitan Bernath's brief. During the Submission Period, we developed the features that directly address Eitan's requirements:

"I Cooked It" tracking — Mark recipes as cooked, track completion, celebrate the moment (Eitan's core ask)
Video recipe import — Paste YouTube, TikTok, or Instagram URLs; we extract the recipe and generate a shopping list (Eitan's exact quote)
RevenueCat integration — Full subscription model with Premium/Family/Lifetime, entitlements, restore purchases
Recipe photos & Photo Journal — "I Made This!" to document what you actually cooked
Gemini AI — Recipe parsing from any URL or video, pantry-aware chat
Demo mode & judge access — Hackathon judges can test all features

See HACKATHON_DEVELOPMENT_TIMELINE.md for commit evidence and a detailed timeline.

The Problem That Inspired Us

As home cooks ourselves, we've all been there: scrolling through TikTok, Instagram, or food blogs, saving recipe after recipe with the best intentions. "This looks amazing!" we think. "I'll make this next week!" But then... life happens. The recipes pile up in our saved folders, forgotten and uncooked.

When we saw Eitan Bernath's brief for the RevenueCat Shipyard Hackathon, it hit home. Eitan perfectly captured the frustration of modern home cooks: we're drowning in recipe inspiration but struggling to turn that inspiration into actual meals. The gap between "I want to make this" and "I cooked it" feels impossibly wide.

We realized the problem isn't lack of inspiration—it's the friction between finding a recipe and actually cooking it. The steps are too many, too disconnected:

Find a recipe (easy)
Save it somewhere (easy)
Plan when to cook it (hard)
Figure out what ingredients you need (hard)
Generate a shopping list (hard)
Actually cook it (hardest)

What We Learned

Through user research and our own experiences, we discovered three key insights:

1. The Shopping List Barrier

The biggest friction point isn't finding recipes—it's the mental overhead of converting a recipe into a shopping list. Most people abandon recipes because they can't easily figure out what to buy.

2. Context Switching Kills Momentum

Jumping between apps (recipe → notes → shopping list → store) breaks the flow. By the time you've copied ingredients into your shopping app, you've lost the motivation to cook.

3. Organization Without Action is Useless

Beautiful recipe collections mean nothing if they don't lead to cooking. We needed to focus on actionability, not just organization.

How We Built Sous

Sous is designed to eliminate every friction point between recipe inspiration and actual cooking. Here's how we built it:

Core Architecture

Tech Stack:

Frontend: React Native (Expo) for iOS and Android
Backend: Next.js with tRPC for type-safe APIs
Database: PostgreSQL with Drizzle ORM
AI: Google Gemini 2.5 Flash for recipe parsing and ingredient recognition
Monetization: RevenueCat SDK for iOS, Stripe for web/Android
Infrastructure: Vercel for hosting, AWS S3 for image storage

Key Features That Solve Eitan's Problem

1. One-Tap Recipe Import 🎯

Problem: Copying recipes manually is tedious
Solution: Paste any recipe URL and Sous automatically:

Parses ingredients, instructions, and metadata
Extracts images and cooking times
Saves everything in one organized place

Technical Challenge: Recipe sites have wildly different HTML structures. We built a robust parser using AI to extract structured data from any format, handling edge cases like embedded videos, complex ingredient lists, and multi-step instructions.

2. Smart Shopping List Generation 🛒

Problem: Converting recipes to shopping lists is manual and error-prone
Solution: With one tap, Sous:

Extracts all ingredients from saved recipes
Groups by category (produce, dairy, meat, etc.)
Generates a shareable shopping list
Tracks what you already have in your pantry

Technical Challenge: Ingredient parsing is surprisingly complex. "2 cups of flour" vs "flour" vs "all-purpose flour" need to be normalized. We built an AI-powered ingredient recognition system that understands variations and quantities.

3. Recipe Discovery from Your Pantry 🔍

Problem: You have ingredients but don't know what to cook
Solution: Sous lets you:

Add ingredients you have
Search recipes across multiple sources (TheMealDB, Epicurious, Delish, NYTCooking)
Find recipes you can make right now

Technical Challenge: Aggregating recipes from multiple APIs with different schemas required careful normalization and caching strategies.

4. AI-Powered Cooking Assistant 🤖

Problem: Recipes don't adapt to your needs
Solution: Sous uses AI to:

Scale recipes up or down
Suggest substitutions
Answer cooking questions
Provide personalized meal planning

Monetization Strategy

We integrated RevenueCat to offer a sustainable freemium model:

Free Tier:

Save up to 10 recipes
Basic shopping lists
Limited recipe imports

Premium Tier ($4.99/month or $49.99/year):

Unlimited recipe saves
Import from any recipe website
Advanced shopping list features
AI-powered meal planning
Recipe scaling and substitutions
Ad-free experience

This model ensures Sous can serve Eitan's audience while building a sustainable business.

Challenges We Faced

1. Recipe Parsing at Scale

Challenge: Every recipe website has different HTML structure. Parsing reliably across thousands of sites seemed impossible.

Solution: We built a hybrid approach:

Pattern matching for common sites (TheMealDB, Epicurious)
AI-powered extraction for unknown sites
Fallback to manual parsing with user-friendly error messages

Learning: Sometimes the best solution is graceful degradation, not perfection.

2. Cross-Platform Monetization

Challenge: iOS requires RevenueCat, Android/web can use Stripe. Managing two payment systems seemed complex.

Solution: We built a unified subscription service layer that abstracts payment providers. The app checks platform and routes to the appropriate SDK, while our backend maintains a single source of truth for subscription status.

Learning: Good architecture means the complexity is hidden from users and most of the codebase.

3. Real-Time Ingredient Recognition

Challenge: Users want to take photos of ingredients and have them recognized instantly.

Solution: We integrated Google Gemini Vision API for image-to-ingredient recognition. The challenge was balancing speed (local processing) vs accuracy (cloud API). We chose accuracy and optimized API calls with smart caching.

Learning: User experience sometimes means choosing the slower but more accurate path.

4. State Management Across Platforms

Challenge: Keeping subscription status, recipe data, and user preferences in sync across iOS, Android, and web.

Solution: We built a centralized backend with tRPC for type-safe APIs, React Query for client-side caching, and careful invalidation strategies.

Learning: Type safety isn't just about preventing bugs—it's about moving faster with confidence.

How Sous Serves Eitan's Audience

The "I Want to Make This" → "I Cooked It" Journey

Before Sous:

See recipe on TikTok → Save to camera roll
Later: Try to remember where you saved it
Find recipe → Manually write ingredients
Go to store → Forget half the ingredients
Get home → Realize you're missing something
Give up → Order takeout

With Sous:

See recipe on TikTok → Copy link → Paste in Sous
Sous auto-generates shopping list
Add to shopping list → Check off at store
Get home → Everything you need is ready
Follow recipe → Actually cook it! 🎉

Specific Features That Address Eitan's Brief

✅ Generate grocery lists from recipe videos or links

Paste any recipe URL, get an instant shopping list
Works with TikTok, Instagram, YouTube, blogs, recipe sites

✅ Organize cooking inspiration

Save recipes from multiple sources in one place
Organize by meal type, cuisine, or custom collections
Never lose a recipe again

✅ Simplify the path from idea to execution

One-tap shopping list generation
Pantry integration (know what you have)
Recipe discovery based on available ingredients
Cooking mode with step-by-step guidance

Technical Highlights

RevenueCat Integration

Proper iOS SDK integration with user identification
Entitlement-based access control
Restore purchases functionality (App Store requirement)
Cross-platform subscription sync

AI-Powered Features

Recipe parsing from any URL format
Ingredient recognition from images
Smart ingredient normalization
Recipe scaling and substitution suggestions

Performance Optimizations

Image caching and optimization
Smart API request batching
Offline recipe access
Optimistic UI updates

What's Next

Sous is just getting started. Future features we're excited about:

Meal Planning: Weekly meal plans with automatic shopping lists
Recipe Scaling: AI-powered portion adjustments
Social Features: Share recipes and cooking achievements
Nutrition Tracking: Automatic nutrition calculation
Cooking Timers: Built-in timers synced with recipe steps

Conclusion

Eitan's brief resonated because it's a problem we've all experienced. Sous isn't just another recipe app—it's a cooking execution system designed to eliminate every barrier between inspiration and action.

We built Sous because we believe home cooks shouldn't have to choose between inspiration and execution. With Sous, every saved recipe becomes a real possibility, and every shopping trip becomes a step toward actually cooking.

For Eitan's audience—home cooks overwhelmed by saved recipes—Sous is the bridge from "I want to make this" to "I cooked it."