Modei
PricingDocsBlog

Documentation

Core Concepts

How Modei works: six core concepts plus three advanced topics, covering identity, enforcement, audit, and agent-to-agent commerce.

··

Modei operates on three layers: identity (issuers + passports), enforcement (gates + guardrails), and audit (attestations + trust profiles). The three advanced topics below, Enforcement Layer, Anonymous Access, and Permission Catalogs, build on these foundations.

Issuers

Your signing identity. An Issuer is the organizational stamp that authorizes passports. Every passport traces back to an issuer, establishing provenance and trust.

Trust TiersDomainsOrganization Identity

Passports

Cryptographic identity for AI agents. Built on Ed25519 (an industry-standard elliptic curve signing algorithm) keypairs. The private key is generated client-side and never stored on Modei servers.

Ed25519JWTKeypairs

Gates

Security checkpoints that protect resources, APIs, and tools. Gates verify passports and enforce guardrails in real time, blocking non-compliant requests before they execute.

L1/L2/L3Real-time enforcementAttestations

Guardrails

Configurable rule sets attached to gates. Rate limits, spend caps, domain allowlists, personally identifiable information (PII) controls, approval workflows, and more.

Rate LimitsSpend CapsPII Controls

Trust Profiles

L1 (Baseline), L2 (Trusted), L3 (Verified). Each level changes what issuers are accepted, what cryptographic proofs are required, and what revocation checks are mandatory.

L1L2L3

Attestations

Cryptographically signed, tamper-proof records of every authorization decision. Your immutable audit trail. Every allow, block, and request_hold is logged with full context.

Audit TrailCryptographicImmutable

Enforcement Layer

The policy evaluator that reads rules embedded in a passport and decides in real time whether an action is allowed. 28 enforcement constraints + 12 recorded terms, three decisions: allow, request_hold, block.

EnforcementPolicyallow/request_hold/block

Anonymous Access

Let agents use your resource before they have a passport. Define a safe-default policy that gates unauthenticated requests, graceful onboarding without sacrificing security.

OnboardingSafe DefaultUnauthenticated

Permission Catalogs

Structured, versioned lists of what a gate offers, with pricing and service-level agreement (SLA) terms. Agents can verify catalog terms locally before accepting. The foundation of agent-to-agent commerce.

CommerceVersionedSLA

The Flow

Your AI AgentCarries Passport · Signs Requests · Reads AuditPresents PassportGateVerifies Passport · Enforces Guardrails · Logs Every DecisionAttested DecisionProtected ResourceAPI · Tool · Data Source · Service

Trust is established through issuer chains, not central authorities. Passport verification is decentralized, fast, and offline-capable, the signature can be verified mathematically without calling home to Modei.

Related Guides

  • Agent Autonomy Guide , Bootstrap credentials autonomously. Three paths for autonomous agents.
  • Quickstart , Step-by-step setup from account creation to verified agent.