The Problem

Before Bitewise, there was no convenient way to derive specific products from something as simple as a picture. If you went out to eat and saw an interesting dish you wanted to make later, or even saw something on social media, you had no practical way of figuring out what went into making it.

The Solution

Enter BiteWise. All the user needs to do is upload a photo, and Bitewise handles the rest. Image bytecode data is first stored in Google Firebase, where the hosted url is then passed into Clarifai's food recognition API. The output of that is refined through Spoonacular's product data API asynchronously to offer specific product recommendations and reference prices which are dynamically summed according to selected products.

What's next?

In the future, as we scale, we could continue to layer API's to retrieve specific products or even train a more granular machine learning model that recognizes brands (or at least more popular ones). In addition, Bitewise would benefit greatly under a partnership with big grocery chains like Walmart or Whole Foods, and could provide even more accurate and relevant data with access to developer and/or internal API's.

More Information

The project is currently available on our Team's Github. Please feel free to email me if you would like to see more.