In the daily struggles of university life, maintaining a balanced and healthy diet becomes increasingly difficult as stresses increase and fast food becomes ever more accessible. Between crammed study sessions and extended sleep deprivation, preparing your own meals becomes a daunting task. CookMe aims to transform meal prep into light work, so you can spend more time on what matters.
What it does
CookMe lets users search through millions of online photos and captions. Users will then be able to select an image of food which piques their interest, after which CookMe will prompt them with several possible recipe search options. They will then be able to select from these recipes to view detailed prep instructions, ingredients lists, and nutritional information about the selected recipes.
How I built it
CookMe is hosted as a web app using React. It gathers data from Facebook's Instagram APIs to supply photos and captions based on specific hashtag search criteria from user-supplied search queries. When the user selects a post, CookMe will feed the associated image into Google Cloud's Vision API, which will generate several (hopefully) food labels for the image. CookMe will then reverse search these labels in a recipe database and present these results to the user.
Challenges I ran into
What I learned
You need at least one front-end developer on a team. And also lots of React
What's next for CookMe!
There are quite a few more user-centric features CookMe can offer, which weren't implemented due to time constraints. For example, we can collect historical search data for each user and attempt to recommend recipes based on a user's preferences., and also take user searches as input to improve the image labeling model.