hungr.ai

“What should we eat for dinner tonight?” Can’t decide? hungr.ai is a web-based application that uses ai to identify what you may be (subconsciously) craving at this moment.

Comment

Inspiration

Choosing what to eat has always been an issue during hangouts. Rather than spending so much time deciding what we want to eat, we wanted to create an app that would tell us using facial recognition!

Basically, we're always hungry but literally never know what to eat.

The idea itself comes from the paper referenced below. We were thinking about if it's possible that we do in fact know what to eat, but on a subconscious level, and as the paper below explains, we do! So the idea of the app is to take subtle facial clues you might give off in reaction to food and see what you truly want to eat, plus it shows a place nearby on google maps for fun!

Pentecost, Claire. “What Did You Eat and When Did You Know It?” Science Journal, 2002, doi:10.1080/00043249.2002.10792126

For the UI, we drew our inspirations from Pinterest and dribble dashboards. We experimented with UI trends (glassmorphism) for fun!

What it does

hungr.ai identifies what types of cuisines or food a user is subconsciously craving using facial recognition & ai.

How we built it

We used a combination for modern platforms, such as figma for design, firebase / cloud functions for the backend, nextjs / react for the frontend.

Challenges we ran into

Our biggest problem was that we quite honestly forgot about the competition for a whole day, meaning that we've been working non-stop for the last 12 hours :sob:

In terms of the UI, we had a hard time implementing glassmorphism because we wanted to use it sparingly to avoid sacrificing the usability, accessibility, and visual hierarchy of our product. We decided to only use glassmorphism as a "card" of our interface, as well as an option to switch to our default material look.

When it comes to development, I'm still implementing facial recognition with the api. I spent way too much time making the app look amazing that our backend is currently lacking. We were originally planning to implement this using google cloud function and a python library, but that went downhill fast because setting it up was too weird, so right now I'm implementing everything in javascript with the help of our buddy tensorflow.

Accomplishments that we're proud of

  • Stunning Design (pats on back)
  • Creating visual hierarchy
  • Technical Complexity within 12 hours

What we learned

  • Design sprinting, how to create fast iterations & prototypes.
  • Using the double diamond design paradigm
  • Designing for desktop view
  • Coding stuff + more!!

What's next for hungr.ai

  • An option to show specific vegan/dietary preferences
  • A mobile application counterpart
  • User testing
  • An option to link users' food delivery accounts to hungr.ai

Built With

Share this project:

Updates