Infrastructure for the agent economy

Every agent.
Verified.

AI agents are starting to buy, book and sign on behalf of businesses. AIN Verify gives each agent a persistent identity, binds it to the organization accountable for it, and lets anyone check — in real time — whether a specific action is within the authority that organization granted.

Model-agnosticEd25519 key-bound identityInstantly revocable authority

AIN / AGENT CREDENTIAL

Agent identity recordEXAMPLE

ATLAS Procurement

AIN-7KQ2-M9XD-4TRA

Principal

Example Corporation

Identity status

● Active

Authorized scope

Purchases · office supplies

Per-action limit

$250.00 USD

IDENTITY • AUTHORITY • AUDIT

01Persistent agent identity

02Principal-owned authority

03Real-time verification

04Instant revocation

The trust stack / 001

Intelligence is abundant.
Trust is infrastructure.

Before an AI agent can purchase, negotiate or execute work in your systems, you need trustworthy answers to three questions.

⌑

Identity

Who is this agent? Every agent gets a permanent AIN identifier bound to an Ed25519 key, independent of the model provider or platform it runs on.

◇

Authority

Who authorized it, and to do what? Organizations issue mandates with action types, counterparties, per-action and cumulative limits, and an expiry — and can revoke them instantly.

↗

Verification

Can this action proceed? One API call returns a deterministic allow/deny decision with a reason code, signed by AIN so it can be kept as evidence.

How it works / 002

Register. Delegate.
Verify.

Register the agent

Your organization registers the agent and receives its AIN. We generate a signing key pair (or you bring your own public key). The private key never leaves the agent.

Issue a mandate

An admin confirms their password and issues bounded authority: what the agent may do, with whom, up to how much, until when. Every change is written to a tamper-evident audit log.

Counterparties verify

Before accepting an order or request, the receiving business calls /v1/verify with the AIN, the action and the agent's signature over that action. AIN returns a signed decision.

curl https://ainverify.com/v1/verify \
  -H "Authorization: Bearer $AIN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "ain": "AIN-7KQ2-M9XD-4TRA",
    "action": {
      "type": "purchase",
      "amount": "125.00",
      "currency": "USD",
      "merchant": "Acme Supply"
    },
    "agent_proof": {
      "timestamp": "2026-10-30T14:02:11Z",
      "nonce": "q8Zr2mVn0cXyP4tL",
      "signature": "3vE1…Ed25519…Qw"
    }
  }'

{
  "object": "verification",
  "id": "ver_9fJ2kLm0QpX8aN1c",
  "decision": "allow",
  "reason_code": "AUTHORIZED",
  "ain": "AIN-7KQ2-M9XD-4TRA",
  "agent": { "name": "ATLAS Procurement", "status": "active" },
  "principal": {
    "name": "Example Corporation",
    "identity_status": "self_attested"
  },
  "agent_proof": { "status": "verified" },
  "mandate": {
    "id": "mnd_…", "per_action_limit": "250.00",
    "remaining_total": "1875.00", "currency": "USD"
  },
  "signature": { "alg": "EdDSA", "kid": "ainv-…", "value": "…" }
}

Public registry

Look up any registered agent's identity and status.

Early access / 003

What AIN verifies today
— and what it doesn't.

Trust infrastructure has to be precise about its own guarantees.

✓ Verified by AIN

The AIN exists and the agent is active, suspended or revoked
The request was signed by the agent's registered Ed25519 key (when a proof is supplied)
The action falls within an unexpired, unrevoked mandate
Every decision is signed by AIN and logged in a hash-chained audit trail

✕ Not yet verified

The legal identity of organizations — it is self-attested during early access and labelled as such in every response
Payments — AIN does not move money or connect to card or bank networks; a decision is an authorization signal, not a payment approval
That a human outside the account approved a mandate beyond password re-confirmation

Architecture / 004

A foundation built to
work everywhere.

01 / REGISTER

AIN Registry

Stable, checksummed identifiers bound to a public key and an accountable organization.

02 / DELEGATE

Mandate Engine

Scoped authority with limits, counterparties and expiry. Revocation takes effect on the next call.

03 / CHECK

Verification API

Action-level decisions with reason codes, proof-of-possession and replay protection.

04 / RECORD

Audit Layer

Append-only, hash-chained history of every identity, authority and decision event.

Pricing / 005

Free during early access.

Paid plans below are proposed and not yet available. We are not collecting payment.

Developer

Free

Everything you need to integrate and test.

Unlimited agents in early access
Mandates and verification API
Audit log and export

Agent Pro PROPOSED

$19 / agent / mo

For developers operating commercial agents.

Production SLAs
Webhooks on revocation
Higher API limits

Enterprise PROPOSED

Custom

For platforms building trusted agent ecosystems.

Verified organization identity
SSO and advanced controls
Dedicated support

Every agent.Verified.

Intelligence is abundant.Trust is infrastructure.