Agentic IAM Solution | Ory

Agentic IAM Overview

Agentic IAM handles authentication and authorization for AI agents, autonomous software that acts on behalf of users or systems to perform tasks, make decisions, and interact with APIs. As AI agents become more prevalent, they need secure identities, scoped permissions, and auditable access trails just like human users.

Ory's Stack for Agentic IAM

Ory's stack supports Agentic IAM patterns:

The key Agentic IAM patterns that Ory supports include agent identity registration, scoped token issuance (limiting what an agent can do), delegation and consent (users authorizing agents to act on their behalf), token chain revocation (instantly revoking an agent's access), and audit logging for compliance and debugging.

Install Ory Agent Plugins

Sign in to your Ory Network project and open /agent-plugins on its canonical *.projects.oryapis.com URL. The page provides installation commands for Claude Code, Codex, Gemini CLI, Cursor, Google Antigravity, GitHub Copilot CLI, Continue, Goose, Cline, OpenClaw, OpenCode, Pi, and Amp. Each command includes the project's canonical URL, the matching Ory Agent Security URL, and the reserved public OAuth2 client ID. The plugins require Node.js 22 or later.

Most plugins install for your user account and apply to every project. The Continue, OpenClaw, OpenCode, and Amp plugins install into the current directory, so run their commands from the project you want to protect.

Why Ory for Agentic IAM?

Every AI agent request—whether from a human, app, or autonomous system—must be verified and authorized with the right permissions. Building secure agent authentication yourself means handling token lifecycles, delegation chains, and revocation across distributed systems. Ory provides the OAuth 2.0 and permissions infrastructure to implement Agentic IAM while maintaining the auditability and control enterprises require.