Fly Quest

Inspiration The release of a fly AI model sparked our curiosity: what could it do with a physical body? We wanted to give it a robot dog’s body to interact with the world and expand its attraction to fruit into the ability to seek out other objects. What it does Fly Quest searches for cans, approaches the user when it finds one, and then guides them to its location. It explores how fly-inspired AI can drive physical navigation and interaction. How we built it We’re building Fly Quest around a fly neural network model and a Freenove robot dog. Our approach connects can detection with the model’s attraction behavior, then translates its output into movement commands. This brings perception, neural processing, and robotic motion together to support the search-and-guide task. Challenges we ran into Accomplishments that we're proud of What we learned What's next for Fly Quest