Backseat Champions

Upgrade. Strategize. Win. Repeat.

Manage your driver’s race in an autobattler with roguelite mechanics, focused on the thrill of racing. Select random upgrades throughout the race, while picking the most optimal tires. Overcome each race's challenges and win co-op championships!

Feel free to try the free demo before buying the game!

• Short-burst races, allowing you to quickly hop in and get straight to the action.

• Every race is unique. Experience different race lengths, dynamic weather conditions and random upgrades.

• All racing is real - nothing is scripted. The drivers are powered by machine learning. Naturally observe the craziest of overtakes and the hardest of crashes.

• Play with or against your friends. Battle it out or compete in the Career Mode championships together.

• Upgrade your car. During the race, select random upgrades, unlock bonuses and build the fastest car to win each race.

• Plan strategic pit stops. Manage your pit stops and utilize many different tire compounds to optimize your race.

• Manage your boost. Manipulate your car by using boosts and aerodynamics buffs. Assist your driver with overtakes, defending and by setting overall faster lap times.

• Navigate the weather. Experience dynamic weather changes throughout the race. Will it stay dry, or risk making a pit stop for wet tires?

• Start your Career Mode. Unlock championships by winning in others. Climb your way through the ladder and become the best ever Backseat Champion!

• Unlock teams and customizations. Level up throughout your racing adventure. Unlock car customization features such as wheel rims, helmet designs, teams to choose from and more!

• Your choices matter. Everyone starts on equal footing. Infinite possibilities. Where will you make the difference?

• Race unique cars. From all kinds of Formula cars, to GT's, and all the way to modern Endurance cars. Face new challenges in different race dynamics.

Note on AI

This project does NOT use Generative AI.

The drivers are powered by reinforcement learning using conventional, non-generative neural network models.

All training was done from scratch using self-made training environments built within the 50+ tracks of the game, and run from personal servers.

No copyrighted material, or third-party data, was used during the development of the game. All assets were created either by commissioned artists or myself.