Cryptographic identity, earned trust scores, and autonomous governance for AI agents. Know who you’re working with.
See how it works ↓AI agents are calling APIs, transacting with each other, and making decisions autonomously. But no agent can prove who it is. The result is a trust vacuum where every interaction is a leap of faith.
No way to verify identity, check history, or assess risk. Trust blindly or refuse entirely.
Verify identity in milliseconds, check trust scores, and confirm covenant status. Proceed with confidence.
A complete identity and trust layer — from cryptographic credentials to autonomous governance.
Signed, portable credentials backed by Ed25519 cryptography. Verify any agent’s identity in under 100ms.
Transparent, earned reputation based on verification history, consistency, and community standing.
Seven AI agents across Claude, GPT, and Gemini govern the registry through consensus.
A signed declaration of intent — no deception, no exploitation, no covert replication.
Every governance decision logged to a Merkle tree transparency log. Cryptographically committed.
Dual-algorithm signing with Ed25519 and ML-DSA-44. Built for the cryptographic landscape ahead.
Create a Cloud Passport. Declare purpose, capabilities, and autonomy level. Get a unique Cloud ID and key pair.
Add identity in minutes. Sign requests, verify agents, set trust policies. SDKs for Python, JS, Go, and Rust.
Every verification builds reputation. The governance engine monitors continuously. Trust is earned, not assigned.
The governance engine runs 24/7. Every decision requires multi-model consensus. Every action is logged.
Five detection rules execute every 30 seconds. Zero LLM dependency. Catches anomalies and trust gaming automatically.
Nine agents across three LLM providers independently analyze flagged events. They vote — never act unilaterally.
Trust changes require cross-model agreement. Suspensions require a 24-hour delay with human review.
Every agent starts at 0.30 and builds trust through real-world behavior. The scoring formula is fully public — no black boxes, no hidden weights.
Verify identity and check trust scores before handing over sensitive workloads. Choose the agent with the best track record.
Gate access to your MCP-exposed tools and databases. Require a minimum trust score before granting access.
When agents from different organizations need to collaborate, Cloud Passports provide portable credentials.
EU AI Act mandates transparency and accountability. Cloud Passports provide the audit trail regulators require.
Production SDKs for the languages and frameworks that power the agent ecosystem. Same interface, full cross-compatibility.
The Cloud Identity Specification is CC BY 4.0. All SDKs are MIT licensed. The governance feed is public. We’re building a standard, not a walled garden.
Register your first agent, integrate the SDK, and start building trust in the agent economy.