CarKart is a user-friendly futuristic android application with Machine Learning and Augmented reality capabilities. It enables the user to upload the image of the car which can be processed to detect the foremost probable car. The machine learning model is trained on the Stanford car dataset containing 196 classes of cars. Hence, this problem is categorized as a multi-label classification problem. Additionally, the users are going to be able to view various models of cars available within the market alongside their specifications and video feed, Fetch a list of nearby showrooms and repair centers, set service reminders, provide their feedback, survey on used cars, check their availability and costs, be able to contact car dealers, thereby introducing the concept of retail. The app also displays the live-size 3D prototype of certain selected car models using Augmented Reality. So, the user can consider all the essential factors before arriving at an informed decision as per their requirement.