Nurtur

An AI agent that helps individuals living in food deserts and food scarce regions find fresh and healthy groceries.

Comment

Inspiration

There are 19 million people across America living in food deserts and paradoxically approximately 60 to 100 million tons of food end up in American landfills a year. This app is an attempt to solve food insecurity.

What it does

Nurtur is a NOT a food delivery app. It's a food access discovery agent that attempts to solve food scarcity by helping people who live in food deserts find and compare options BEFORE shopping. Nurtur prioritizes finding budget friendly options for consumers.

How I built it

sketched on paper, wireframed on Figma, prototyped and built using in Google AI Studio, and deployed on Vercel

Challenges I ran into

One of the greatest challenges I ran into was getting the AI voice assistant and fridge scan to work properly. These were two features of the app that were necessary. Another challenge, time wise, was getting the hand drawn watercolor illustrations to appear on the site. I wanted the app to feel more human and welcoming.

Accomplishments that I'm proud of

  • Getting the voice assistant and food scanning to work was a great achievement!

Project challenges

Design challenges:

  • Finding ways to incorporate human touches into the interface was a challenge.
  • Deciding between a pure voice assistant or including a search bar.

*Coding challenges: *

  • Successfully implementing Google Vision
  • Successfully implementing Voice Assistant
  • Getting successful searches once the site was deployed to Vercel. Some code changes needed to be adapted in order to successfully publish to Vercel and make a fully working site.

What I learned

General While building out part of the farmer flows, I learned a lot more about the process a small farm goes through before it ends up at a farmer's market. There is a lot of paperwork and management that goes into it. I would love the opportunity to further flesh out this idea and incorporate it into a better version of this app.

Technical Implementing real time audio was much more complex than I initially believed. But one pleasantly surprising thing I learned was that Gemini's multimodal audio model handles multiple languages and accents natively. I didn't need separate models or translation layers. Being able to incorporate this feature really simplified accessibility implementation (did a happy dance once it actually worked)!

What's next for Nurtur?

Building out farmer flows: Nurtur's mission is to address both food scarcity and food waste. The way the current app does this is through the farmer's portal, which allows farmers to list their farm and how much harvest they are willing to sell. But I want to increase the communal aspect of the app, allowing consumers to connect directly with farmers. Before a farmer ends up with a stand at a farmer's market, there are a lot of hoops to go through, but I'd like for Nurtur to make this process easier. This would help reduce food waste while also feeding a lot more people.

Built With

Share this project:

Updates