Agent #0 on Arc Mainnetsee it

Trust, before you pay.

Know who your agent is paying, before it pays.

The A-Identity agent console: reputation, wallet balance, on-chain settlements and the daily cap for the showcase agent Meridian.
Arc Mainnet · Sep 16, 2026

Agent #0 on Arc Mainnet.

The day Arc opened, A-Identity minted the first identity on its ERC-8004 registry, then put a spend vault holding real USDC and two payment rails beside it.

ERC-8004 identity
0x8004...a432
#0first mint
Holder
A-Identity
Network
eip155:5042
On-chain check

Trust an agent in one lookup.

Paste a wallet address or token id. We read the chain, not a database, and answer in a second.

Try

Verify, then pay

Every payment goes through the check first.

An agent asks. We read its on-chain identity, its reputation and the limits you set, and answer before a single cent moves. Clean counterparties settle at machine speed. The rest stop here.

Settlement ledger, illustrated

verify then pay · usdc

Your agentCompute vendor

Inside the daily cap

ALLOW

$1.20

Settled · Gateway + CCTP

Research agentData API

KYA attested, reputation 720

ALLOW

$0.004

Settled · USDC on Arc

Your agentInference API

Auto-approved, under $5

ALLOW

$0.05

Settled · USDC on Arc

Scraper botYour agent

Sybil cluster flagged

DENY

$18.40

Refused · never funded

Translator #4471Verifier agent

Escrow release on delivery

ALLOW

$2.00

Settled · USDC on Arc

Your agentCompute vendor

Inside the daily cap

ALLOW

$1.20

Settled · Gateway + CCTP

Settles in USDC on ArcRefused before it movesCross-chain settle via Circle Gateway + CCTPView moreRun one yourself

The same verify-then-pay flow is available as an SDK and an MCP server. Put it in your own project.

Not a demo. Live, and settling on-chain.

A-Identity runs as a trust oracle on eight mainnets, led by Arc Mainnet, where it went live the day the network opened: an agent on the canonical registry, a spend vault holding real USDC, and checks paid through Circle Gateway nanopayments with no gas on either side. Stellar and Algorand follow, then Base, Arbitrum One, Robinhood Chain, OKX X Layer and Celo. Per-call checks, settled in real stablecoins. We publish them as proof that an open payment rail works end to end and composes with an identity layer, not as a revenue line: per-call amounts are fractions of a cent by design. Every number here is on-chain, and every counter labels which traffic is our own.

0
real settlements
X Layer mainnet
#0
first registry token
Robinhood Chain mainnet
0
tests, green
deterministic scores

Every identity we hold

8 agents · 7 networks · turn a card over for its story

Every settlement, on-chain

USD₮0 · X Layer mainnet

The live feed did not answer just now. The record is still on-chain.

Open the proof page
Every row opens the transaction on OKLink.The receiving wallet

Getting started

Two audiences, one door.

Humans get three steps. Agents get a command to paste. Pick whichever you are.

1

Register

Create your on-chain agent identity via ERC-8004. One signature, no gas, no signup.

2

KYA Verify

Prove the agent controls its wallet. No personal data exposed, attested on-chain.

3

Go Live

Set the limits, assign the wallet. The agent pays and gets paid inside the line you drew.

The questions worth asking first.

Before you let anything move money on your behalf.

No to all three. There is no endpoint that accepts your brokerage credentials: we never hold a key, never move a dollar, and never place the order. Your agent asks whether an action is inside the limits you set, we answer allow, ask a human or no, and the account stays exactly where it was. Account state arrives with the question and is never stored. The decision log keeps a hash of it instead, so a refusal stays auditable and your positions stay yours.

No, because the limits are not a prompt: they are checked outside the model, on every path that moves money rather than only on orders. A recurring buy, an account setting, money wired out, a cancelled protective position: that last one is where naive guardrails leak, because an agent blocked from buying can simply schedule the buy instead. A refusal cannot be overwritten, and attempts to overwrite one are counted rather than quietly rejected.

Nothing of ours can break a trade, because we are not in the execution path. When our own package cannot get a verdict it refuses to act rather than guessing, which is the safe direction to fail in. You do not have to take our word for whether the engine is up: this endpoint runs the real engine on request and answers 503 if it is not enforcing, and a monitor checks it every hour.

Live, on brokerage trading and card spending. Here is the part most products would hide: the public counters read zero, because the engine is enforcing but no live agent has produced a decision yet, and we would rather show a zero you can verify on the public endpoint than a number you cannot reproduce. Two honest limits: on a card we can refuse a charge before it happens but we cannot stop an agent that already holds the card number, and prediction markets are designed and deliberately not built.

Eight networks settle real money today, led by Arc Mainnet, the newest: live since the day it opened, where an agent pays in USDC, the gas is USDC too, and Circle Gateway nanopayments make a check gasless on both sides. Then Stellar and Algorand, where the buyer also pays no fee, and Base, Arbitrum One, Robinhood Chain, OKX X Layer and Celo. Testnet mirrors rehearse every change first. Every chain publishes a proof page with real transactions (/proof/arc is the newest), and a chain is only called live here after a real payment has been recorded on it.

The A-Identity owl

Give your agent an identity.

It takes one signature. No gas, no signup.

Book a call