Astrix Security is an identity security product for AI agents and non-human identities such as service accounts, OAuth tokens, machine credentials and secrets. It builds a real-time inventory of agents, MCP servers and NHIs with context on risk and business use, flags excessive privileges and abnormal activity, and provisions new agents through its Agent Control Plane with short-lived, just-in-time, precisely scoped credentials. Astrix is now part of Cisco and ended standalone sales of new licenses effective June 30, 2026, while existing customers continue to receive service and support.
CISOs, security architects, IAM engineers and cloud security teams in large enterprises are the people it was built for, especially those who found that employees were granting agents broad, static permissions outside normal access reviews. The company cites customers such as Workday, HubSpot, Boomi, BigID, Workato and Mercury. Compliance and risk teams use its policy and audit capabilities, and its academy offers a certification in agentic security for practitioners who want to specialize.
Typical work includes discovering which agents and service accounts have access to corporate assets, pruning overprivileged ones, and setting policy so new agents are created secure by design. Agentic threat detection and response spots compromised credentials and out-of-scope agent actions, and non-human ITDR watches for suspicious machine identity activity and third-party breaches. Lifecycle management follows each agent and NHI from provisioning to decommissioning. Centralized secret management covers vaults and cloud environments, and third-party risk management assesses outside apps and vendors that touch the environment.
