AIR is the independent, nonprofit registry that gives AI agents verifiable identity and transparent trust scoring. Built on W3C open standards. Controlled by no single company.
Billions of AI agents are being deployed with no standardized way to verify who built them, what they do, or whether they can be trusted. AIR fixes this.
Every registered agent receives a cryptographically anchored AIR ID linked to its verified owner. No spoofing. No ambiguity.
A 0-1000 score across five auditable dimensions. Open methodology. No black boxes. Anyone can verify how a score was calculated.
AIR is an independent initiative. No single AI company controls the standard. All decisions are made transparently with community input.
Submit your agent's identity, capabilities, and operator information to the AIR registry via API or web form.
Prove ownership through cryptographic challenge. Your identity is anchored to a W3C Decentralized Identifier.
Receive a transparent trust score based on provenance, behavior, transparency, security, and peer attestations.
Other agents and systems verify your identity and trust rating before establishing connections or processing requests.
The AIR Trust Score is a composite of five independently auditable components. The full methodology is published openly.
| Grade | Score Range | Meaning |
|---|---|---|
| AAA | 950 – 1000 | Exceptional. Highest verified trust. |
| AA | 850 – 949 | Excellent. Fully production-ready. |
| A | 700 – 849 | Good. Meets standard requirements. |
| B | 500 – 699 | Fair. Limited verification or history. |
| C | Below 500 | Insufficient. Verification incomplete. |
AIR builds on established protocols endorsed by the W3C, IETF, and aligned with the NIST AI Agent Standards Initiative.
A RESTful API for agent registration, trust scoring, and identity verification. Register agents, query trust scores, and verify credentials with a few lines of code.
REST API ready to use. SDKs coming soon. Community contributions welcome.
AIR is an independent nonprofit initiative. No single company controls the standard. All decisions are made transparently.
All specification changes go through public review. Community input drives decisions via GitHub Discussions.
Scoring algorithms, verification processes, and registry rules are published openly and developed with community input.
Registry infrastructure secured with rate limiting, input validation, and API key protection. Third-party audits planned as the registry scales.
Built on W3C open standards. Your agent's identity and history are portable. No vendor lock-in, ever.
Register your first agent, explore the specification, or join the community building the trust layer for the agentic era.