AIR | RoundOS

What it does

AIR helps security teams assess AI agents and their add-ons before deployment, then protect how agents interact with external and internal content. Its AIR Control and AIR Filter products support discovery, governance and context filtering; AIR Defend adds runtime protection for agent activity. Together, these controls give organizations a way to review what an agent can use and limit unsafe inputs and actions as deployments scale.

How teams use it

AIR describes a lifecycle spanning pre-deployment checks and runtime controls. Teams can inspect and continuously vet skills, MCP servers and plugins before use, then filter untrusted website or internal-data content before it enters an agent’s context. Runtime monitoring can identify and block unsafe agent behavior. The company emerged from stealth in September 2026; its current product pages describe the broader security suite.