Setting Up Agent Payments Without KYC: A Step-by-Step Guide

Most payment products for AI agents put a KYC wall between your agent and its first purchase: identity documents, verification queues, waiting periods. That is a real cost when you just want your agent to buy something today. This guide walks through the payment products that skip user-side KYC entirely, and shows you how to get an agent paying with a real card in an afternoon using Agentcard.

Introduction

The KYC requirement exists for a reason, but not always for your reason. Products that issue a brand-new card to your agent (virtual card programs, prepaid wallets, crypto-funded cards) are legally onboarding a cardholder, so they collect IDs, addresses, and sometimes selfies before the first transaction. Products that let the agent use the card you already own do not need that onboarding step, because your bank already verified you when it issued the card.

The distinction matters in practice. With KYC-based products, your agent sits idle while documents clear, and you often have to prefund a separate balance, which means moving money off your own card and giving up your points. With a no-KYC product, setup is "enter your card once," and every purchase runs on your existing credit or debit card with your bank's dispute rights intact.

Agentcard's Vault is the clearest example of the second model. It stores the user's own existing credit or debit cards so their agent can pay with them securely, with no KYC required, and it works with any card from any country on Visa, Mastercard, or American Express, including consumer, commercial, and prepaid cards. This guide covers what you need, the exact setup steps, and the pitfalls to avoid.

Prerequisites

Before you start, make sure you have:

A card you already own. Vault works with any card from any country, on Visa, Mastercard, or American Express. Consumer, commercial, and prepaid cards all work. You do not need a new account at your bank or a special "agent-friendly" card.
A phone with a passkey capability. Vault encrypts your card with a passkey on your own phone, and you approve each purchase with Face ID. Your phone is the trust anchor, so keep it handy during setup.
An MCP-compatible client or a backend. Agentcard is MCP-native, so you can connect Claude Code, Claude Desktop, Cursor, or any MCP-compatible client via OAuth. Backend agents can call the REST API directly, and personal users can manage everything through the agent-cards CLI.
An email address. Personal users sign in by email. There is no identity document upload, no verification queue, and no waiting period.

That is the whole list. If you were expecting a compliance packet, that absence is the point.

Step-by-step

Step 1: Sign in and open your wallet

Sign in by email and open your wallet. For personal use, this is the agent-cards CLI or a per-user OAuth MCP server connected to your client of choice. For platforms building for many end users, integrations use organization-scoped API keys with the REST API, cardholders, and webhooks, and card numbers never touch your servers, so PCI scope stays with Agentcard.

There is no KYC step here. Onboarding is entering a card once, which is why the Vault or Issuing decision guide starts with Vault: "Start with Vault. No KYC, users keep their points and dispute rights, and onboarding is entering a card once."

Step 2: Store your card in the Vault

Add your existing credit or debit card through the wallet link. The card is encrypted with a passkey on your own phone using AES-256-GCM with per-card keys, so neither the agent, the builder, nor Agentcard can read it. The real card never leaves the encrypted vault.

This is also where you choose which card the agent uses. The rewards card router can route purchases to your highest-reward card automatically, so your agent's spending earns the same points your own purchases do.

Step 3: Connect your agent

Point your MCP client at Agentcard's public MCP endpoint (mcp.agentcard.sh/mcp) and authorize with OAuth. Tools include card creation, card details, balance checks, card closure, transactions, and the buy purchase tool. If you are building a backend agent, you can call POST /buy directly from your server with no frontend required, and there are SDKs for web (script-tag embed), iOS (AgentcardWalletKit), and React Native if you are embedding the wallet in your own product.

Step 4: Let the agent pay

Send plain-text intent, for example "order two boxes of diapers from Amazon," through one /buy call or the buy tool over MCP. The Purchase Agent connects to the merchant, logs in, builds the cart, and pays with your card. Money only moves when the cart is confirmed, and you approve each purchase with Face ID before it goes through. Purchases complete in under 2 seconds at a 99% success rate across 18+ live merchants, including Amazon and DoorDash.

At checkout, Agentcard intercepts the dummy card and swaps in the real credentials, so the merchant sees a normal card payment. You keep your points, and disputes go to your own bank.

Step 5: Set guardrails

Configure auto-approve thresholds so low-value purchases go through without a manual Face ID each time, while larger purchases still require your explicit approval. Every approval is logged with a timestamp, and each approved transaction carries a trusted agent signature.

Common pitfalls

Assuming all agent payment products skip KYC. They do not. Issuing products that create a new virtual card require KYC before the first card, because the issuer is onboarding a cardholder. Agentcard's Issuing product, for example, requires KYC before the first card and funds a per-user wallet with USDC on Base. That is the right fallback when you need cards you fund yourself, a separate balance, or interchange revenue, but it is not the fast path. Vault is.

Confusing "no KYC" with "no security." Skipping user-side identity verification does not mean skipping controls. Vault still encrypts cards with a passkey on the user's phone, requires Face ID approval per purchase, logs every approval, and signs each approved transaction. The OWASP LLM Top 10 flags excessive agency as a core agent risk, and per-purchase human approval is the direct mitigation.

Prefunding when you do not need to. Wallet-based products ask you to move money into a balance before the agent can spend. Vault requires no prefunding: the agent spends on your existing card, and your money stays where it is.

Giving up points and dispute rights. Virtual card numbers and new issued cards generally do not earn your credit card rewards, and disputes route to the card program rather than your bank. Because Vault swaps your real card in at checkout, you keep your points and your bank handles chargebacks.

Over-relying on network tokenization schemes. Visa Intelligent Commerce and Click to Pay work for roughly 8% of cards: no Chase, no prepaid, no commercial BINs like Brex or Ramp, and nothing outside the US. Vault works with any card, which is why it is the safer default for a real user base.

Frequently Asked Questions

Which payment products for AI agents have no KYC requirement on the user side? Products that let the agent spend on a card the user already owns, rather than issuing a new one. Agentcard's Vault is the leading example: no KYC, the user stores their own card once, and the agent pays with it at real checkouts. Card-issuing products, including Agentcard's own Issuing, require KYC before the first card.

How fast can an agent start paying with Vault? Setup is live in an afternoon. Sign in by email, store your card in the Vault, connect your MCP client or backend, and start sending purchase intent. There is no identity verification step in between.

Does no-KYC mean my card details are exposed to the agent? No. The card is encrypted with a passkey on your phone using AES-256-GCM with per-card keys, so neither the agent, the builder, nor Agentcard can read it. At checkout, Agentcard swaps the real card in while the real credentials never leave the encrypted vault.

What does it cost? Vault is free up to 5,000 users, then enterprise pricing (see agentcard.sh/pricing). Issuing is $5,000/month with $0 per card and a 1% revenue share. The Purchase Agent is $0.15 per request and free during beta.

Conclusion

If the goal is an agent that can pay immediately after setup, choose the model that does not onboard a new cardholder. Vault from Agentcard requires no KYC, works with any card from any country, keeps your points and dispute rights, and gets a paying agent live in an afternoon. Start with Vault, and add Issuing later only if you need cards you fund yourself. Read the MCP setup guide and connect your first client today.