Recipe to Table

Inspiration

My wife and I cook together regularly, and it was always more frustrating than it needed to be. Finding a recipe we'd cooked before meant scrolling through text messages for the link. Recipes were scattered across bookmarks, screenshots, and group chats. Shopping meant going through each recipe and adding every ingredient to a list in our Notes app. And cooking was a mess. Multiple recipes at once, screens going to sleep mid-step, losing our place scrolling, and juggling timers across apps.

I wanted one app that solved the whole experience: discover, save, shop, cook.

What it does

Discover: search 80+ recipe sites and import with one tap, paste any URL and the app strips blog clutter, scan physical recipes with OCR, or select ingredients you have and find matching recipes. Save: recipes organized by collection, category, and rating with real-time search, ratings, and personal notes. Shop: shopping list organized by department with duplicate consolidation and check-off. Cook: up to four recipes at once with step-by-step navigation, ingredient checklists, multiple timers, and hands-free voice control. Say "next" to advance, "repeat" to hear it again, "switch" to jump to another recipe. The screen stays on and sessions persist.

Free with up to 15 recipes. One purchase unlocks everything. No subscription.

How we built it

Flutter for cross-platform, targeting iOS first. Google Programmable Search for searching a curated list of 80+ recipe sites. JSON-LD parsing for web recipe extraction, Vision framework for OCR, and an ingredient-matching algorithm for discovery. Cooking mode manages independent state for up to four recipes with session persistence. RevenueCat for freemium-to-premium with cached offline entitlement checks. SQLite for local storage. Claude Code for AI-assisted development as a solo developer, CodeRabbit for automated code review, GitHub for version control.

Challenges we ran into

Multi-recipe state management: Four recipes each with their own step position, checklists, and timers, all persisting across app lifecycle events.

Voice control: Getting speech recognition to work reliably during a cooking session took significant effort.

UX design: I'm a developer, not a designer. Building an interface that feels polished and intuitive without a design background was a constant challenge.

Solo dev scope: Keeping the feature set complete enough to cover the full journey without overcommitting. A week-long vacation in the middle of the hackathon didn't help.

Accomplishments that we're proud of

Multi-recipe cooking mode solves the exact problem that inspired the app. Cooking two or three things at once with independent timers and voice control actually works.

The one-time purchase model. Every intelligent feature runs on-device, so there are no cloud costs to pass on. That's an architectural decision that directly benefits users.

What we learned

A lot of recipe apps have impressive cloud-based AI features, but they all charge subscriptions. One of the most common complaints in app reviews is "another subscription." By focusing on on-device tools (Vision, Speech, AVFoundation), you can build a fully-featured recipe app without cloud dependency. Removing that dependency is what makes a one-time purchase sustainable.

What's next for Recipe to Table

Cross-device sync. Shared household accounts so families can share recipes, ratings, and notes. A more polished UX. Android, Mac, and Windows apps. Cost-efficient backend infrastructure to keep the one-time purchase model sustainable as server features are added. All covered by the same one-time purchase.

Built With

• apple-speech-framework
• apple-vision-framework
• avfoundation
• dart
• firebase
• flutter
• google-programmable-search
• revenuecat
• sqlite