The Farama Foundation is a 501(c)(3) nonprofit that advances reinforcement learning through open-source tools and shared standards for researchers and industry practitioners.
We develop and maintain widely used core libraries that provide standard APIs across many kinds of reinforcement-learning environments, including robotics simulations and video games. We also maintain many popular environments used by the research community. You can view the full list of our libraries on our projects page. Collectively, they have more than 400 million installations and are used by tens of thousands of projects, making them some of the most widely installed libraries in reinforcement learning and machine learning.
Our day-to-day work focuses on making major quality-of-life, productivity, and performance improvements to our core libraries; bringing important RL environments up to our standards; keeping them compatible with our core libraries; and supporting them with maintenance and bug fixes. We intend to make research workflows faster, more reliable, and easier to reproduce.
The Farama Foundation is led by 12 core members responsible for projects and key areas of the organization, and is supported by a global community of more than 800 contributors. Our core contributors bring experience from research, engineering, academia, and industry, and many have been active in the reinforcement learning community for years.
Our community is open to anyone interested in reinforcement learning, and we welcome contributions and donations. Our community is primarily organized on Discord and through our GitHub organization. You can also donate to help fund full-time engineering work on these libraries. For funding inquiries or other questions, contact us at contact@farama.org.