Draft for public comment · v0.1 · March 2026

The open standard for
AI agent identity

AIS-1 defines a bonded identity pair — one card for the agent, one card for the legal entity or person responsible for it. Together they enable identity, compliance, and legal standing for the agent economy.

Read the specification → Submit feedback
CC0 — no rights reserved
ERC-721 · ERC-8002 · W3C DID
Hedera HCS canonical log
Open for comment until 30 June 2026

The Standard

What is AIS-1?

AIS-1 defines an open smart contract standard for a bonded identity pair — a single cryptographic token that permanently links an AI agent's identity to the legal entity or person responsible for it. Neither party can be separated from the other without revoking the bond.

The standard addresses what we call the Wild Agent Problem: the near-total absence of formal identity, legal accountability, and compliance infrastructure for the hundreds of millions of AI agents currently active globally. India's national Doot / Agent One programme — which proposes a personal AI agent for every citizen, bound to their Aadhaar identity — is one of the first national-scale demonstrations of why bonded agent identity matters.

Agent Card
Agent DID (did:ais1:…)
Agent name & type
Capabilities array
Chain addresses
AML status
Model / framework
Metadata URI (IPFS)
BOND
Sponsor Card
Sponsor DID
Legal name
Entity type
Jurisdiction
Registration no.
KYC status
Verifiable Credential URI
Bond hash: keccak256(agentDid ++ sponsorDid ++ timestamp ++ tier)
HCS log: Hedera Consensus Service — immutable
Soulbound: ERC-8002 — non-transferable
Without AIS-1
Agents have no identity — unverifiable, unaccountable
No responsible party when an agent causes harm
AML/KYC frameworks cannot apply to agent transactions
FATF Travel Rule cannot travel with agent payments
Agents cannot legally hold assets or enter contracts
With AIS-1
Every AI agent has a unique, verifiable DID
Sponsor permanently bonded — accountability is cryptographic
verifyBond() performs CDD in milliseconds at every transaction
Bond ID travels with transaction — Travel Rule satisfied natively
Sovereign tier provides government-attested sponsor credentials and physical data anchoring

Tiers

Three tiers. One ladder.

A single mandatory standard would fail — forcing full KYC on a prototype agent achieves nothing. AIS-1 defines three tiers: enter at Basic with no friction, graduate to Verified when commercial stakes require it, achieve Sovereign status when enhanced due diligence and government-attested credentials are required.

TIER 0
Basic
Permissionless. Anyone can issue. No KYC/AML required. Self-declared sponsor. Gas fees only. Suitable for development and prototype agents.
Permissionless
TIER 1
Verified
Authorised issuer only. The issuer checks who the sponsor is — identity, beneficial ownership, sanctions screening — and confirms the agent is cleared for commercial use. That clearance is embedded directly in the bond as a machine-readable credential, readable by any counterparty in milliseconds.
Identity checked · AML cleared
TIER 2
Sovereign
Enhanced tier. The sponsor undergoes enhanced due diligence including beneficial ownership verification and source of funds review. The sponsor card carries a government-attested Verifiable Credential — a cryptographically signed credential from a recognised authority confirming the sponsor's legal existence. A physical data presence in the issuing jurisdiction anchors the bond off-chain.
Enhanced · Government-attested

Verification

One function. Full compliance. Milliseconds.

Any counterparty calls verifyBond(bondId) before transacting with an agent. It returns everything needed for AML/KYC compliance in a single on-chain call.

// Counterparty verification flow — before any agent transaction

const result = await ais1.verifyBond(bondId);

// Returns:
//  valid      bool    — is this bond active and not revoked?
//  tier       uint8   — 0=Basic · 1=Verified · 2=Sovereign
//  sponsorDid string  — DID of the accountable sponsor
//  amlStatus  uint8   — 0=unverified · 1=cleared · 2=suspended

if (result.valid && result.tier >= 1 && result.amlStatus === 1) {
  // Agent is Verified, AML-cleared, accountable
  // Proceed with transaction
} else {
  // Reject — agent not cleared for commerce
}

This is customer due diligence at agent-commerce speeds — performed once by the authorised issuer, readable by every counterparty in milliseconds, for the lifetime of the bond. KYC done once. Compliance everywhere.


Timeline

From working paper to deployed standard.

MAR 2026
AIS-1 v0.1 published
Specification and companion document published on GitHub under CC0. Open for public comment. First reference implementation deployed on Base Sepolia testnet.
LIVE
APR 2026
First bond issuances
The first two AIS-1 bonds issued: Bond 1 — an AI agent bonded to a company sponsor, and Bond 2 — an AI agent bonded to an individual person. Both bond IDs published on Agent-Pulse, becoming the first verified entries in the world agent census.
IN PROGRESS
MAY 2026
Consensus Miami — live demo
Live AIS-1 bond issuance at EasyA × CoinDesk Consensus 2026, Miami. Contract running on Base testnet. Hedera HCS canonical log integration demonstrated live.
MAY 5–7
Q3 2026
AIS-1 v1.0 mainnet
Production deployment on Base mainnet. AIS-1 Basic and Verified tiers operational. DID resolver submitted to W3C Universal Resolver. Bond IDs published on Agent-Pulse — the first verified entries in the live agent census.
Q3 2026
Q4 2026
Sovereign tier + multi-chain expansion
Sovereign tier specification finalised. AIS-1 deployed on Ethereum and Arbitrum in addition to Base. Hedera native implementation. Sovereign tier becomes operational as qualifying jurisdictions emerge.
Q4 2026
2027
Agent commerce integration
AIS-1 integration guides published for MPP and x402 payment protocols — showing developers how to call verifyBond() before processing agent payments. Agent-Pulse live integration tracking verified vs unverified agent population in real time.
2027

Documents

Specification and supporting materials.

Specification
AIS-1 Specification v0.1
The full technical specification. Definitions, dual card structure, tiers, smart contract interface, DID method, security considerations, implementation roadmap.
Published · GitHub
Smart Contract
Reference Implementation
AIS1.sol — deployable Solidity contract. ERC-721 + ERC-8002 soulbound. Full issuance, revocation, verification and recovery functions. OpenZeppelin security patterns.
Published · GitHub
DID Method
did:ais1 Specification
W3C DID Core conformant method specification. CRUD operations, DID Document structure, resolution algorithm, security and privacy considerations.
Published · GitHub
JSON-LD Context
AIS-1 Vocabulary v1
Machine-readable vocabulary defining all AIS-1 attributes. W3C JSON-LD 1.1 conformant. Enables semantic web interoperability for agent identity data.
Published · GitHub

Public Comment

AIS-1 v0.1 is a draft for public comment.

Feedback is invited from AI agent developers, blockchain engineers, legal and regulatory professionals, enterprise AI deployers, standards organisations and government bodies. The comment period for v0.1 closes 30 June 2026. A revised v0.2 will incorporate substantive feedback.

We are looking for:

Technical review of the smart contract interface and soulbound implementation · Feedback on the tier structure and KYC/AML architecture · Input from regulatory bodies on FATF compliance mapping · Review from W3C DID Working Group on the did:ais1 method · Proposals for authorised issuer status · Real-world deployment use cases and edge cases · Support for formal standardisation through IEEE, ISO or IETF