Proxectos usando SuperTuxKart

    Due to the libre and open source nature of SuperTuxKart, the game is used in various research project, studies or simply prototypes made by fans

    On this page, you can discover various projects that have been done with our game.

    If you have a project using SuperTuxKart not listed here, or if you want to use the game in your own project, you can contact us.

    Contidos

    Driver Development

    Tyr

    SuperTuxKart was used as a target for the development of Tyr, a driver for the Mali GPUs found in ARM SoCs, porting the C code of Panthor to Rust.

    In November 2025, Tyr’s ability to render SuperTuxKart races at low settings with reasonable performance for the tested SoC was showcased in Collabora’s blog post about Tyr’s progress as well as through a video showing STK played with the new driver.

    Technical Demonstrations

    A Mercedes as SuperTuxKart Controller

    When it comes to driving a car in the real world, the big touchscreens that have become commonplace in modern cars are often distracting. And a real car is a tad expensive to use as a gaming device.

    Nonetheless, when Mercedes-Benz showcased SuperTuxKart being played with a car’s controls at the 2019 Mobile World Congress, journalists were wowed.

    Many outlets referred to ‘Mario Kart’ in their reporting, probably because any kart game is ‘Mario Kart’ to some, but it’s SuperTuxKart which powered this impressive demo. And perhaps this carelessness in naming could be seen as a testament to SuperTuxKart quality.

    STK’s open nature further allowed Mercedes to synchronize the car’s lighting, air conditioning and even seatbelt with in-game events, making the experience exceptional.

    Microsoft IllumiRoom Demo

    IllumiRoom was a proof-of-concept system from Microsoft Research.

    It used projected visualizations around a television screen to enhance the traditional living room entertainment experience.

    Showcased in 2013, it featured SuperTuxKart 0.8 in its demo video, with white circles projected on the wall representing falling snow.

    A screenshot from Microsoft Illumiroom with Mozilla from SuperTuxKart 0.8 driving in Northern Resort and snow falling effect on the walls surrounding the TV screen

    Illumiroom with SuperTuxKart on the TV

    Virtual Joystick using Kinect

    The Gesture Therapy platform is a system for upper limb rehabilitation developed by a Mexican team.

    To showcase the capabilities of the Virtual Joystick system it uses, they released a demo video showing the acceleration and steering of a kart in SuperTuxKart 0.8 being controlled through the movements of an arm.

    Academic Research and Teaching

    A Reduction of Imitation Learning and Structured Prediction to No-Regret Online Learning

    A landmark paper from 2010 by Stéphane Ross & al introducing the ‘DAgger’ method for machine learning, improving the methods of imitation learning. This paper is still referenced well over a decade later. It used SuperTuxKart as a core part of its experiments.

    The full paper is available as well as an old video showing the simple machine learning models attempting to drive.

    Other Machine Learning Projects

    SuperTuxKart is used regularly by people learning machine learning at universities. Here are some example of such projects:

    Game-Based Rehabilitation for Myoelectric Prosthesis Control

    An Austrian study from 2017 evaluating the use of video games for the rehabilitation of upper-limb amputees. As, according to the study, many ‘prosthesis users abandon their devices due to difficulties in prosthesis control’, it explored games as a way to make the training process more enjoyable and efficient.

    SuperTuxKart is one of the three games that was used as part of the study.

    Comparing the Effectiveness and Ergonomics of Smartphone-Based Gamepads

    A 2024 study from the University of Regensburg setting 24 people to drive in SuperTuxKart, comparing both their objective performance and subjective experience with different smartphone-based gamepads as well as a physical gamepad.

    The detailed methodology and results highlighted in the full paper are very interesting.

    Many participants felt that the absence of shoulder triggers was the biggest limitation of the smartphone-based gamepads (and the STK team agrees).

    Smartphones with shoulder triggers are now available on the market. Through this study, SuperTuxKart is playing a small role in making them more common!

    A figure from a scientific paper comparing SuperTuxKart round times between a physical gamepad and smartphone-based gamepads

    The physical gamepad proved noticeably better, despite input latency being equalized between the tested devices.

    Service-oriented Integration of SuperTuxKart

    A paper from 2022 proposing the use of a REST API to make it easier for services to interface with the game, and presenting a possible implementation.

    Concepts for Advanced Integration of SuperTuxKart into Connected Cars

    A detailed master’s thesis from the University of Stuttgart published in 2021 exploring how a connected car could be used to play games. The full text is freely available.

    A Study on QoE Improvement of Online Games with UDP Multipathization by SDN

    A paper from 2020 from the Nagoya Institute of Technology exploring the use of multipathization for online multiplayer. SuperTuxKart is one of the two games used in the study.

    Teaching video game translation: first steps, systems and hands-on experience

    A Brazilian paper from 2018 published in Texto Livre Linguagem e Tecnologia, with the full research paper freely available online. In section 3 ‘Why choose SuperTuxKart?’, the authors explain how SuperTuxKart’s open nature made it a natural fit for didactical purposes.

    Localization in the Era of Inclusivity

    A master’s thesis from the University of Geneva published in 2024, available in French.

    Towards More Accurate Open Source Kart Race Balancing

    A paper from 2014 from the Central Michigan University comparing the powerup distribution between SuperTuxKart 0.8, Super Mario Kart, and Mario Kart 7.

    Fun projects

    Controlling a STK kart with hand gestures

    A fun project from 2024 used hand gestures and chair orientation to control a kart. A script emulated a joystick using the data from a vision recognition system identifying gestures and from a phone on the player’s chair which determined the turning angle depending on the chair’s orientation.

    The driving level possible with this control method is surprisingly good considering how funny it looks!

    A SuperTuxKart window and a camera feed where gestures are analyzed to drive a kart.

    Open Street Map

    What about driving on the streets you know in SuperTuxKart? This is what a project from 2013 wanted to make possible, with a method using OpenStreetMap data and importing real world data explained on the project’s page.

    As it requires the use of the now very old SuperTuxKart 0.7.3 and also a fair amount of manual work for good results, it is not very practical, but it shows how SuperTuxKart is a game uniquely suited for cool experiments.