In April, we launched Gym⁠, a toolkit for developing and comparing reinforcement learning⁠ (RL) algorithms. With Universe, any program can be turned into a Gym environment. Universe works by automatically launching the program behind a VNC⁠(opens in a new window) remote desktop—it doesn’t need special access to program internals, source code, or bot APIs.

Today’s release consists of a thousand environments⁠(opens in a new window) including Flash games⁠(opens in a new window), browser tasks⁠(opens in a new window), and games like slither.io⁠(opens in a new window) and GTA V⁠(opens in a new window). Hundreds of these are ready for reinforcement learning, and almost all can be freely run with the universe⁠(opens in a new window) Python library as follows:

The sample code above will start your AI playing the Dusk Drive Flash game. Your AI will be given frames like the above 60 times per second. You’ll need to have Docker and universe installed.

Our goal⁠ is to develop a single AI agent that can flexibly apply its past experience on Universe environments to quickly master unfamiliar, difficult environments, which would be a major step towards general intelligence. There are many ways to help⁠: giving us permission on your games, training agents across Universe tasks, (soon) integrating new games, or (soon) playing the games.