Made by Beyond MVPs
Make what you crave
reci-p.ai is an app for Android and iOS that provides users with the ability to quickly and easily recreate recipes of their favourite shop-bought food items at the tap of a button. Users can take or upload photos of their food and receive great tasting recipes that can be made at home from fresh. Recipes can be saved and favourited, and viewed across devices. We wanted users to get their recipes in as few steps as possible, with a clean and easy to understand UI. As such, the app also integrates Google Sign-in so users can get started as quickly as possible, and leverages Google Gemini to provide an accurate and quick response.
The Team
Simon Busby
I’m a former print designer turned developer with a…
background in creating clean, beautiful content, keen to combine my coding skills and design expertise to build engaging, well-crafted digital products.
Lauren Evans
Junior software developer with an interest in assistive…
technology and statistical tools.
Samin Taseen
Recent biomedical sciences graduate from the University of…
Oxford. Now transitioning into software development, I am keen to combine my scientific background and coding skills in a tech environment.
Florin Patroescu
Curious and dedicated junior software developer with a…
background in logistics and administration, now fully committed to building a career in technology
Tech Stack
We used: React Native, Expo, MongoDB, Express, Nodejs, Jest, Supertest, Render, and Gemini We wanted a native app feel that would work for both iOS and Android, so we opted for React Native with Expo for our front-end. For our back-end we used a MongoDB database to store our users and recipes. MongoDB also allowed us to store images without relying on a separate cloud-based storage solution. We used Express and Nodejs to create our RESTful API, and for testing, we used jest and supertest. We used Expo and React Native packages to handle OAuth with Google, image picking, camera functionality and text extraction. Doing so kept up with our goal of making sure our app was cross-platform. Text extraction significantly reduced our overhead for calls to Gemini 2.5 flash, the AI API that we are using to generate a structured output of the things needed to give the user a meaningful recipe.
Challenges Faced
We had issues with some of the packages that could have been resolved with more effective spiking. The original OCR engine we planned on using was web only, and the alternative package we used was broken on Android. We had some issues with our original implementation of Google Sign-in which was resolved with a more up-to-date package. Otherwise, the only large issue that we faced was getting iOS builds running on physical devices due to the paywalled nature of the Apple Developer Program. We enjoyed learning new technologies like React Native and MongoDB and delving deeper into styling the application.