Everything you need to know.

What is A-Identity?

A-Identity is an identity and payment layer for AI agents. It gives every agent:

• A verified identity
• Clear permissions
• A payment wallet
• A record of its actions

Why do agents need identity?

AI agents can already use apps, APIs, and digital services. But many cannot prove who they are or who controls them. Without identity, trust is weak. Without trust, agent payments cannot grow safely.

What does KYA mean?

KYA means Know Your Agent. It checks:

• Who the agent is
• Who operates it
• What it can do
• Which limits it must follow

Is KYA the same as KYC?

No. KYC verifies a person or company. KYA verifies an AI agent, its owner, and its permissions.

What is ERC-8004?

ERC-8004 is a standard for agent identity, reputation, and validation. It helps systems recognize an agent and review its past activity. It provides proof, but it does not guarantee that every action is safe.

Does verification mean an agent is safe?

Not always. Identity proves who the agent is. Permissions decide what it may do. Policy checks decide whether one action should be allowed.

Who sets the limits?

The agent's owner or organization. Limits may control:

• Spending amount
• Approved recipients
• Allowed tools
• Daily activity
• Human approval
• Blocked actions

Can an agent change these rules?

No. The rules are checked outside the AI model. The agent cannot change them through conversation or prompts.

What decisions can the system return?

There are three possible results:

• Allow: the action is within the rules.
• Ask: a person must approve it.
• Block: the action is not allowed.

When is human approval needed?

Human approval may be required when:

• The amount is high
• The recipient is new
• The action is unusual
• Money leaves an approved account
• A setting is changed
• The action is hard to reverse

The agent can work quickly, but people keep control of important decisions.

How do agents pay each other?

A-Identity uses x402 for agent payments. The agent receives a payment request, proves its identity, passes the policy check, and completes the payment.

What does x402 do?

x402 handles the payment request and response. A-Identity adds the trust around it. It checks:

• Who is paying
• Who is receiving
• Whether the amount is allowed
• Whether approval is needed

Does A-Identity make money from x402?

Not in any amount that matters, and we do not present it that way. The settlements on our proof pages are fractions of a cent each. They are published as evidence that an open payment rail works end to end and composes with an identity layer, which is the thing we are building. Across the market, x402 volume is growing while revenue stays small, so we treat it as infrastructure to build on rather than as a business line.

What does "Web2 trust, Web3 rails" mean?

The product should feel simple and familiar. Behind the interface, blockchain systems handle payments and settlement. Users see clear rules. The infrastructure works in the background.

Which asset is used?

USDC is the main settlement asset. Other stablecoins may be added when the selected network and payment system support them.

What role does Arc play?

Arc provides the blockchain infrastructure. It supports USDC based fees and fast transaction finality. A-Identity uses Arc to connect agent identity, payment, and settlement in one flow.

Is Arc live on mainnet?

Arc is currently available as a public testnet. Any Arc activity should be described as testnet activity until mainnet is available.

Is A-Identity only a wallet?

No. A wallet moves money. A-Identity also checks identity, permissions, risk, and approval.

Does A-Identity hold my money?

No. Funds remain in the user's or company's own account or wallet.

Does A-Identity store private keys?

No. Private keys stay with the selected wallet or account provider.

Does A-Identity need my account password?

No. It does not need brokerage passwords, wallet recovery phrases, or full account credentials.

Does A-Identity place transactions?

A-Identity checks whether an action is allowed. The connected wallet, broker, or payment provider completes the action.

Does verification expose private data?

Only the required information should be shared. The goal is to prove a fact without revealing the full record behind it.

Does A-Identity use zero knowledge proofs?

Zero knowledge verification is part of the roadmap. It may allow an agent to prove that it is approved or trusted without sharing all private data.

Does A-Identity store my portfolio?

The system may read the information needed for one policy check. The full portfolio should not be stored. A cryptographic hash can be kept to show which data was used for the decision.

What is stored in the decision log?

A record may include:

• Agent identity
• Requested action
• Decision
• Time
• Policy version
• Reason for the result

Private keys and full account credentials are never included.

What happens if the system is unavailable?

The action stops. If the system cannot return a clear decision, the agent should not continue. It fails in the safer direction.

Can an agent bypass a blocked action?

No. A blocked action cannot be changed by the agent. Attempts to avoid the rule can also be recorded.

Are indirect actions checked?

Yes. The system can also check:

• Scheduled payments
• Recurring purchases
• New recipients
• Account changes
• Transfers
• Cancelled safety orders

The same rules apply even when the agent tries a different path.

Who is A-Identity for?

A-Identity is built for:

• AI agent developers
• Fintech products
• Agent marketplaces
• Automation platforms
• Enterprise systems
• AI commerce products

It is useful wherever agents need to prove, act, and pay.

How can developers connect it?

Developers can use:

• An SDK
• An MCP server
• A provider adapter

What is MCP?

MCP helps AI applications connect to tools and services. It provides the connection. A-Identity provides identity, permission, and payment control around that connection.

Is the MCP server read only?

The first version is read only. Agents can check information and rules without receiving permission to move money.

Can new payment providers be added?

Yes. A new wallet, broker, or payment service can be connected through an adapter. The core policy system does not need to be rebuilt each time.

Is A-Identity live?

The policy engine and available prototype integrations are active. Arc based activity is currently on testnet. The site separates clearly between:

• Live features
• Testnet features
• Demo features
• Planned features

Why are public counters at zero?

Because no public production decision has been recorded yet. A real zero is better than a number that cannot be verified.

What does the system cover today?

The current policy scope includes:

• Brokerage trading
• Card spending checks

More services can be added through adapters.

Are prediction markets supported?

Not yet. They require separate rules, integrations, and regulatory review.

Do you hold my keys, move my money, or see my portfolio?

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.

Can the agent talk its way past the limits?

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.

What happens if you go down while my agent is running?

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.

Is this live, and what does it cover today?

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.

Which chains can my agent pay on today?

Seven networks settle real money today: OKX X Layer, Robinhood Chain, Arbitrum One, Base, Celo, Stellar and, newest, Algorand, where x402 v2 payments are gasless for the buyer: it signs a fee-zero USDC transfer and the pooled group fee covers it. Two testnet mirrors rehearse every change first. Every chain publishes a proof page with real transactions (/proof/algorand is the newest), and a chain is only called live here after a real payment has been recorded on it.