A Guth Labs publication ↗
Back to News

Oleria previews AI Gateway to control agent actions

Green glass gateway controlling a gold pathway between three metallic nodes.
AI-generated editorial illustration

Oleria announced an AI Gateway preview and expanded its identity-governance platform to cover AI agents and the credentials they use, according to its company release. The expanded platform is available to customers, while the gateway is scheduled for general availability in December 2026.

The gateway sits between agents and applications accessed through Model Context Protocol and REST interfaces. Oleria says it holds credentials outside the agent and can permit an action, block it or send it to a person for approval.

The broader update combines discovery of human accounts, service accounts, API keys and agents with ownership assignment and access reviews. It also maps permissions that an agent obtains through another identity, rather than treating the agent as an isolated account. Ownership can be reassigned when an employee leaves.

Oleria says the platform uses both granted permissions and observed usage to inform decisions. These are vendor-described capabilities, not independently tested security guarantees.

For companies deploying agents, the distinction matters: reviewing an account's permissions and checking an agent's individual actions address different parts of the same access-control problem.

Sources and citations

Revision history

Published .