When discussing ideas, we toss around the idea of playing "Go" with people except with a giant map and geolocation to get people moving around. However, we thought that "Go" was too strategic. We needed something more "chaotic" and "freeform."

Dungeons and Dragons came to mind for a highly imaginative and organic experience. One of our team members brought up the idea of an AI-controlled Dungeon Experience similar to AI-Dungeon.

So, Geoverse was born to get people moving around and experience unique stories generated by AI Machine Learning and Natural Language Processing.

A Global Health Crisis

  • Globally, about one in three people get little, if any, physical activity. [1]
  • Roughly two out of three U.S. adults are overweight or obese, and one out of three is obese. [2]
  • At the current rate, more than 1 billion adults are projected to be obese by 2030. [3]

What it does

  • A first-of-its-kind fitness Natural Language Processing mobile game that generates completely unique storylines.
  • Hybrid text-adventure and exercise phone game where you have to walk or run to progress through an ever-evolving AI-generated text-adventure.
  • You are in control of both your fitness goals and your Geoverse avatar!
  • Get healthy in a smart and fun way while sharing your unique adventure with your friends.

How we built it

  • We used Flutter for the front end
  • We used Open AI's GPT3 model for the Natural Language Processing
  • We used Google Cloud for the fitness API and OAuth

Challenges we ran into

Since Google API and Flutter use HTTP version 2.7 and Open AI uses HTTP version 2.9, we couldn't use both Google fitness API and Open AI GPT3 simultaneously. However, we came up with a way around it by using an express server from Flutter to Open AI GPT3. Unfortunately, due to limited time, we couldn't implement both at the same time.

Accomplishments that we're proud of

  • We successfully used the GPT3 model to create a story
  • Our Frontend development team used Flutter for the first time

What we learned

  • We learned how to make applications in Flutter since our frontend team had previously never used Flutter.
  • We learned industry-standard project management- thanks to Zing.

What's next for Geoverse

  • Multiplayer support with random encounters and interactions with other users
  • Ability to set fitness levels and custom goals
  • Better AI story generation
  • Leaderboard and statistics to track your fitness advancement over time
  • Gradual ramping up of fitness goals
  • Advanced fitness exercises like pushups, skipping, etc
  • AI to detect advanced fitness exercises

Use of Google Cloud

We use Google OAuth to access Goole Cloud's Fitness API, which is used to calculate the distance a user walks. We can use this data to verify the amount of distance traveled; hence, we verify if users have completed tasks.


  1. World Health Organization. Notes for the media: New physical activity guidance can help reduce risk of breast, colon cancers; 2011. Accessed January 28, 2012.
  2. Flegal KM, Carroll MD, Kit BK, Ogden CL. Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999-2010. JAMA. 2012;307:491-7.
  3. Kelly T, Yang W, Chen CS, Reynolds K, He J. Global burden of obesity in 2005 and projections to 2030. Int J Obes (Lond). 2008;32:1431-7.

Try it out Links

We have 3 links:

  1. Frontend with Flutter
  2. Backend with GPT3, Google Fitness API, and Django
  3. UI/UX design on Figma
Share this project: