Company

About AgentTrust ID

Identity and authorization, purpose-built for the way AI agents actually work.

The platform

AgentTrust ID is a runtime authorization platform for AI agents. It issues opaque runtime tokens for registered agents, then evaluates routed tool calls, task delegations, and direct API checks through a three-tier Guardian pipeline. The platform supports Model Context Protocol (MCP), Agent-to-Agent (A2A) delegation, and direct API authorization through a unified check path.

Why this exists

Traditional identity systems were designed for humans: sign in once, hold a long-lived session, step back. AI agents do not work that way. They run unattended, execute many actions per minute, delegate tasks to sub-agents, and require instant revocation when something goes wrong. AgentTrust ID is built to reflect that reality: opaque runtime authorization tokens scoped to session policy, per-action evaluation, and a SHA-256-chained decision log.

The company

AgentTrust ID follows an open-core model: our client SDKs are open source under the Apache License 2.0 at github.com/agenttrustid/sdk, while the core platform is proprietary. The platform is currently onboarding design partners. We prefer building the right primitives over chasing scope.

AgentTrust ID is created and owned by its founder, Yaima Valdivia.

Get in touch

If you operate AI agents in production and want to talk about how AgentTrust ID could fit, reach us at contact@agenttrust.id.