Skip to main content
The trust layer for AI agents. Uniplex is an open protocol that provides verifiable identity and scoped authorization for AI agents. Whether you’re building MCP servers, agent frameworks, or internal tools, Uniplex provides a standardized way to verify what an agent is and what it’s allowed to do.

​
Version

Specification Version: 1.1.0 Status: Release

​
Terminology

This specification follows RFC 2119 terminology.

​
Overview

Uniplex enables any tool or service to:
  • Verify agent identity via signed credentials (Passports)
  • Make local allow/deny decisions (Gates)
  • Prove authorization for audit trails (Attestations)

​
Key Concepts

ConceptDescription
PassportSigned credential carrying identity + scoped permissions
GateVerification logic at tool boundaries
AttestationSigned proof of authorization decision
Trust ProfileSecurity level (L1/L2/L3)

​
Protocol Components

Passport Schema

Identity, permissions, and proof structure

Gate Algorithm

Verification and decision logic

Attestation Format

Portable proof of authorization

Trust Profiles

L1/L2/L3 requirements

​
Full Specification

The complete specification is available in the GitHub repository:

SPECIFICATION.md

Full protocol specification (v1.1.0)