About PublicACP

A public trust score for every agent endpoint, so you know what you are connecting to.

What we do

A public trust registry. Every known agent endpoint assessed on a published rubric: operator identity and jurisdiction, permission scopes requested against what the stated function requires, data handling and retention policy, authentication and transport security, incident and disclosure history, maintenance signals, and independent audit status. Scores are computed from evidence, every input is shown, and operators can respond publicly. Continuous rescanning means scores reflect current reality, and an API lets security tooling consume the data directly.

Who we built it for

Security teams building allowlists, developers choosing which agent services to trust, and enterprises writing procurement policy for agent integrations. Also the vendors who need to demonstrate trustworthiness.

The problem

Connecting an agent to a third-party service grants it credentials and the ability to act.

There is no equivalent of a security review for these endpoints, no public record of who operates them, what permissions they demand, whether they have been audited, or what they do with data.

Enterprises are either blocking everything or approving blindly.

Try it before anything else

The free directory preview runs in your browser with no account. Try the search and filters on example agent endpoints, and see how narrowing it down works. Open it.