Aten Security, founded in 2025 and based in San Francisco, specializes in enhancing security for artificial intelligence agents through its flagship product, Thoth. Thoth addresses a critical need in enterprise environments: controlling high-risk actions performed by AI agents post-authorization. The company recognizes that authorization alone is insufficient for compliance and has developed a behavioral enforcement infrastructure that learns from real production data.

Their MOSES behavioral engine, operational since August 2025, classifies agent behaviors and flags deviations to ensure that AI actions align with intended policies. With features like shadow mode, Thoth allows security teams to observe agent actions without blocking them, enhancing their confidence before implementation.

Aten Security's commitment to AI governance is driven by a mission to fill the security gap that many enterprise tools overlook. As a member of the AARM Foundation, Aten promises compliance with industry standards and provides runtime enforcement solutions that are unobtrusive yet effective, delivering decisions in under 100 milliseconds.