Inspiration
What it does
An app to make suggestions for recipes based on an unstructured natural language query provided by the user. We can determine food choices through a highly scalable layer of Watson NLP applied before searching for recipes. The app can currently determine whether someone wants similar foods or different foods from current offerings. It can also suggest food based on ethnicity, occasion, ingredients, and a wide host of other criteria.
How we built it
We made a native iOS app using Swift, and connected it to the Watson Conversation API with a node.js backend, and then used its intent analysis and the original food query to use the Spoonacular API to suggest recipes.
Challenges we ran into
We struggled greatly in setting up the server for the node.js backend, and had to make HTTP GET requests which were glitching.
Accomplishments that we're proud of
The backend use of Natural Language Processing to greatly expand the number of situations we can offer recipes for.
What we learned
How to use artificial intelligence for innocous causes, and create server-side RESTful APIs usable by a frontend through a simple interface.
What's next for FoodBite
An integrated Alexa app and an integrated Android app
Log in or sign up for Devpost to join the conversation.