The goal of this project is to have an agent play in a game of Line Runner via Minecraft. The agent is set to learn to play through computer vision. We will be conducting visual computation through a linear regression classifier to implement object recognition in our agent. This was the approach of our choice because traditional object recognition methods would have a difficult time in a world populated solely by squares. We will be implementing computer vision to allow the agent to see the incoming obstacles and compute the appropriate actions (moving left, right, or continue forward). The agent will be tested through various levels of difficulty. We plan to thoroughly explore open cv, external libraries, and other possible applications of taking screenshots in the code, all while testing our agent against our oracle.