Livia

Find your perfect match in the city of tomorrow.

Comment

Inspiration

Nexhack's page described the theme of this hackathon was going to be called "Turing City where AGI has already been achieve" It got us wondering if AGI has already been achieved what more can we make, dev tools sounded boring, AI B2B SaaS sounded boring and some idea of simulation clicked in my head. We decided that we wanted to use the theme of "Turing City" and shared this concept with my friend which clicked immediately as he was recently watching a Black Mirror episode of this futurstic city where people dated each other in a virtual world overlooked by a super intelligent being that orchestrates these events.

What it does

We built Turing City, a virtual city where autonomous versions(AI avatars) of people form relationships on their behalf. A user signs up using ChatGPT, allowing the system to learn from their past conversations, communication style, emotions, and personality traits to create a detailed personal profile. That profile becomes a virtual character living inside Turing City, where it autonomously meets and dates other users’ characters, progressing through multiple relationship stages and simulated conversations to discover genuine compatibility over time. The system observes these interactions and identifies the most balanced and stable match. When a strong match is found, it automatically schedules a real world date and notifies both users with only the other person’s name and meeting details so the first time they truly meet is in person.

How we built it

For the web application we used Typescript and Nextjs for language & framework and found an opensource repository on building the virtual isometric city that represents Turing City. We also have an iOS app made using Swift. We used to Overshoot's API as the "orchestrator" that tracks our character's movements within the city. Also a special thanks to Cursor.

Accomplishments that we're proud of

With 24 hours we were able to take our idea into a working concept that with a little refinement can be deployed and marketed for users to join start dates. We have also implemented a predictions market that allows people to bet on the outcomes of these virtual dates.

What we learned

We had a very ambitious idea and only 24 hours to build, so we learned to split tasks tactically and prioritize features that would lead us to a working concept that we can demo.

What's next for Livia

Fix the bugs to improve the user experience to be smooth and deploy this for public use. Our goal is to have city simulation that people can view all around the world while still keeping privacy in my mind by giving the users a random identity to keep all the user anonymous without losing the function of the program.

Built With

Share this project:

Updates