MagicBite

Your Lists
Your Pantry
Recipe Instruction
Recipe Instruction (2)
Add From Camera
Revenue Cat Implementation
Recipe Detail
Home

Inspiration

I started cooking about eight months ago. It was my first time preparing food for myself and my loved ones, and as someone who had never touched a pot before, I suddenly found myself spending a lot of time around cooking content. Like many people today, my feed filled up with recipes, creators, and ideas I wanted to try.

The problem was that inspiration lived in one place, and real life happened somewhere else.

Videos were saved but never structured. Ingredients weren’t tracked, and I often forgot what I already had in my pantry, which led to food waste. Shopping became guesswork. And cooking apps often felt cold, cluttered, or overly monetized, not something you genuinely enjoy using.

Part of my personal vision board this year was building things. I wanted to create something that helps people and me cook in a way that feels guided, playful, and frictionless, like having a helper beside you rather than another tool demanding attention.

MagicBite is my answer to that.

What it does

Gather all your recipes in one app, whether from social media or a cookbook
Follow recipes step by step
Keep your kitchen in sync with your recipes
Generate shopping lists directly from recipes
Enjoy music and other apps while following a recipe
Save recipes in different languages and translate them to your native language

How we built it

My go-to tool for building mobile projects has always been React Native. Over the last two years, the ecosystem has grown enormously, sometimes even intimidating, but also incredibly exciting. I wanted to explore as much of it as possible with this project.

MagicBite is built with Expo (SDK 54) using the bare workflow. I enjoy having full control over the development process, and this approach unlocked the native capabilities I needed to bring my ideas to life.

For the backend, I’m using Bun with Elysia. It was my first time working with this stack, but given my Node.js background, it felt like a great opportunity to experiment. Luckily, one of my best friends helped me build several endpoints along the way. I also had some experience with Puppeteer, which I used for scraping content before. Finally, and of course, AI plays a key role in processing posts and videos.

Accomplishments that we’re proud of

I’m trying to become more confident in myself. I never thought I’d reach this point. I didn’t even have a developer account before ShipYard, and paying for one from Venezuela was a complete mess, so the first thing I’m proud of is delivering a working build in this short time.

Following organizers’ advice, I started sharing my process on Instagram. I still don’t feel ready for X yet — but I’ll get there. Seeing friends and early testers genuinely excited about the first versions of the app and saying they’d actually use it meant a lot.

I’m also proud of every line of code behind this project. As a seasoned mobile developer, I pushed the limits of my skills and architected an entire app from scratch just by myself.

What we learned

It might sound strange, but I had never used Expo in a production project before. All the apps I’ve built were brownfield React Native setups. For this project, I wanted to use libraries that required Expo, so I treated it as an opportunity to fully learn the ecosystem.

I had never implemented in-app purchases in the applications I’ve worked on before — usually there’s a dedicated team handling payments — and I was always a bit intimidated by the responsibility of dealing with money. However, working with RevenueCat made the process surprisingly approachable. Through this experience, I learned a great deal about monetization, and it’s now something I feel confident implementing and fully understand going forward.

What’s next for MagicBite

Building MagicBite solo was something I set out to do this year, and Shipyard felt like the signal to fully commit. This is just the beginning — and there’s a lot I want to explore next.

Making Cooking Fit Real Life

I want recipes to exist beyond the app itself:

Widgets for step tracking without opening the app (and maybe experimenting with Voltra)
More multitasking-friendly experiences
Further reducing friction between cooking and everyday phone use

Expanding Human-Centered Features

I want MagicBite to serve different types of users:

Gym-focused nutrition workflows
Allergy-aware assistance
Health-oriented tracking
Deeper personalization

And eventually a stronger premium tier with planning and calorie systems.

Community-Powered Shopping Intelligence

One direction I’m especially excited about:

Receipt scanning
Crowd-built store databases
Incentives for user contribution
Smarter purchasing decisions

This could grow into its own ecosystem.

AI Pipeline Evolution

Technically, I want to push further:

Extract multiple recipes from single videos — many creators pack several into one post
Improve understanding of creator content
Expand automation capabilities