Trust layer for AI agents
Verified installs, enforced permissions, and cryptographic proof — so platforms and users know exactly what an agent can do before it runs.
How it works
Notary
An independent notary reviews your agent declaration, signs an Ed25519 Trust Card, and issues a verifiable badge.
Trust CenterShield
Runtime policies enforce network access, secret handling, and tool risk. Outcomes are allow, block, or needs_approval.
View agentsPortal
A directory of verified agents, Trust Card API endpoints, and embeddable widgets — ready for any platform.
Embed docsWho it's for
Operators
Teams deploying agents into production.
- Runtime policy enforcement and approval gates
- Shield score and tier gating
- Incident store and audit trail
Publishers
Agent builders shipping to marketplaces.
- Signed ProofCard with Ed25519 badge
- Notary attestation and tlog entry
- Publisher onboarding via GitHub issue
Partners
Platforms embedding trust signals.
- Embeddable trust widget under 5 KB
- Trust API with JSON and SVG endpoints
- HMAC partner adapter and directory API
See it live
Install-ready agents
Verified bundles you can download, unzip, and run locally — with enforced permissions from the first request.
Ready to integrate?
Get your agent verified, earn a ProofCard, and ship with a badge your users can trust.