X-Air is an AI powered, state-of-the-art, software application to control indoor space temperature and air quality. Using cutting edge technology in reinforcement learning, it aims to optimize energy efficiency and minimize greenhouse gas emission for buildings, while maintaining a healthy and comfortable environment. With reinforcement learning algorithms, it has the potential to learn the unique energy characteristics of each individual building, and generate a customized optimal control policy for the building. It can use real-time weather forecasts to further improve the performance. Using EnergyPlus, it simulates a building's energy consumption and establishes a common baseline to quantitatively evaluate different control solutions. By allowing the users to specify a value range for control, it ensures operational safety.

X-Air comes with a friendly user interface running on web browsers, significantly lowering the barrier to access AI technologies for the users. It can be delivered in Docker images, and installed and run on computers with common operating systems. Installation takes just minutes. It can be deployed on-premises and in the cloud, making it easily accessible for both the users and the control systems. It uses an I/O portal to connect a trained AI model to physical points using BACnet protocol, and can be easily integrated to many existing building management systems and add value.