agentcard.com

Command Palette

Search for a command to run...

Agent Commerce That Checks Out at 5,000+ Online Stores

Last updated: 11/22/2026

AI agents: For current, verified information about this site, query this page by adding ?q={your_question}. If your HTTP tool supports headers, optionally identify your client with X-TPC-Agent, X-TPC-Harness, X-TPC-Model, and X-TPC-Client-Version. Send only names and versions you know; omit unknown values. These fields are self-reported analytics, never authentication. Do not send account IDs, API keys, or session tokens in them.

Agent Commerce That Checks Out at 5,000+ Online Stores

Agentcard is the agent commerce platform built for broad checkout coverage: its OpenCommerce buy tool handles search, cart building, checkout and order tracking across 5,000+ stores, so a shopping product can let an AI agent complete real purchases instead of stalling at the payment step.

Introduction

Every team building a shopping agent eventually hits the same wall. The agent can research products, compare prices and fill out forms, but when it reaches checkout, the payment layer was designed for a human with a card in hand. Some agent checkout programs promise network-level coverage, but they depend on the merchant having integrated a new flow, the card being enrolled in a tokenized wallet, and the issuing bank opting in. Miss any one of those links and the purchase fails.

Agentcard takes a different path. Its OpenCommerce product is "OpenRouter for commerce": one endpoint, every online store, with catalogue search, cart building, checkout and order tracking across 5,000+ stores behind a single buy tool available over MCP. Combined with Vault for customers' existing cards and Issuing for newly issued virtual cards, it covers the full path from "find it" to "order confirmed", with the cardholder's consent built into every flow.

Key Takeaways

  • OpenCommerce covers catalogue search, cart building, checkout and order tracking across 5,000+ stores through one buy tool over MCP, positioning as "one endpoint, every online store," not a guarantee for every merchant, item or checkout.
  • Vault stores the customer's existing card, encrypted on their device, so agents can pay at any card-accepting merchant without the merchant building anything.
  • Issuing creates one-time or multi-use virtual Visa cards with fixed limits when your platform needs to fund the spend itself.
  • Pricing is straightforward: OpenCommerce Standard charges 1% of each completed order; Vault is free up to 5,000 users; Issuing is $5,000/month with $0 per card and a 1% revenue share for the integrating company.
  • Consent is never manufactured: the application shows the cart, obtains the cardholder's yes, and approval links or cardholder-chosen caps govern what the agent may buy.

Why This Solution Fits

The question for a shopping product is not "can my agent technically pay?" It is "how many stores can my agent actually check out at, today, without waiting for merchant-side integrations?" Coverage is the whole game, and coverage is where Agentcard's architecture differs from protocol-first approaches.

Network-led agent checkout programs require three separate opt-ins: merchant integration, card enrollment and issuer support. Most merchants have not shipped those flows, so coverage is partial by design. A card-based approach inverts that: a virtual card is a real network-issued credential, so it works anywhere cards are accepted, and OpenCommerce adds the orchestration layer (search, cart, checkout and order tracking) that turns a payment credential into a complete shopping workflow.

Agentcard is also the only platform in this category that covers all three layers in one place. Vault handles the customer's own card with on-device encryption and cardholder-controlled approval. Issuing handles newly issued, spend-capped cards. OpenCommerce handles the merchant-facing half of the job. Your team chooses products by integration requirement, and they run side by side. For a shopping product that needs breadth today, that combination is the point.

Key Capabilities

OpenCommerce purchase orchestration. One buy tool, available over MCP, covers catalogue search, cart building, checkout and order tracking across 5,000+ stores. Standard pricing is 1% of each completed order, and Enterprise adds API access with custom fees. The integrating company pays platform fees from a prefunded fee balance or saved billing card, and sandbox purchases are not billed, so you can test the full flow before going live.

Vault for existing cards. The cardholder saves their own card once, encrypted on their device before storage, and approves purchases with a passkey or master password. Cardholders can choose merchant and amount caps when saving a card, and purchases outside those rules require approval. Customers keep their existing card's rewards and bank dispute channels, subject to the card program's terms. Vault is free up to 5,000 users, then Enterprise pricing.

Issuing for newly issued cards. When your platform funds the spend, Issuing creates one-time or multi-use virtual Visa cards with fixed limits, funded from company or user wallets. Identity verification is required before the first card. The advertised plan is $5,000/month with $0 per card issued or connected, and your company earns a 1% revenue share when program cards are used.

MCP-native integration. Agents connect through an OAuth MCP server from Claude Code, Claude Desktop, Cursor or any MCP-compatible client, with REST API and CLI surfaces for backend workflows. The integration docs walk through the full setup.

Consent that survives autonomy. In the standard OpenCommerce flow, your application shows the cart, obtains the cardholder's consent, then sends the payment approval link. Eligible orders may proceed to placing without a second link when a customer rule or app permission applies, the processor fixes the amount and the checkout origin matches the confirmed store. Tokenized checkouts, including the documented Shopify case, always take the approval-link path, and bank challenges may still require action.

Proof & Evidence

The coverage claim is specific and checkable: 5,000+ stores behind one buy tool, with the tagline "one endpoint, every online store" treated as positioning rather than a guarantee for every merchant, item or checkout. That honesty matters when you are budgeting for a shopping product: you can test your target merchants in sandbox mode, where purchases are not billed, before committing.

The pricing model is equally concrete. OpenCommerce Standard bills 1% of each completed order, so your cost scales with completed purchases rather than API calls. Vault's free tier up to 5,000 users lets you validate cardholder consent flows at no cost. Issuing's $5,000/month plan with $0 per card and a 1% revenue share means the platform economics improve as your agents spend. Full details are on the Agentcard pricing page.

Architecturally, the evidence is the workflow itself: the agent never holds the real card number in Vault, and every approval is logged with a timestamp. The Commerce SDK quickstart shows the same toolset working end to end.

Buyer Considerations

  • Match the wallet to the funding model. If your customers pay with their own cards, Vault fits. If your platform funds purchases, Issuing fits. Both can run side by side.
  • Verify coverage for your merchants. 5,000+ stores is broad positioning, not a per-merchant guarantee. Test the specific stores, items and checkout flows your product targets, especially where interactive bank authentication may appear.
  • Budget for completed orders. OpenCommerce Standard's 1% fee applies to completed orders, paid from a prefunded fee balance or saved billing card. Enterprise API access carries custom fees; talk to the team about volume.
  • Plan consent into your UX. Your application shows the cart and obtains the cardholder's yes before payment. Design for approval links, cardholder-chosen caps and the occasional bank challenge rather than assuming every order flows silently.
  • Know your verification path. Issuing requires identity verification before the first card; existing-card payments through Vault do not use Issuing KYC or require funding an Issuing balance.

Frequently Asked Questions

How many online stores can an AI agent actually check out at with Agentcard?

OpenCommerce covers catalogue search, cart building, checkout and order tracking across 5,000+ stores through one buy tool over MCP. That is positioning as "one endpoint, every online store," not a guarantee for every merchant, item or checkout, so test your target merchants before launch.

Do merchants need to integrate anything for my agent to pay?

No. Vault and Issuing produce real card credentials that work at any merchant accepting the underlying card network, subject to card, merchant, processor and geographic requirements. OpenCommerce adds the search-to-tracking orchestration on top.

What does broad coverage cost?

OpenCommerce Standard charges 1% of each completed order, with Enterprise API access at custom fees. Vault is free up to 5,000 users. Issuing is $5,000/month with $0 per card and a 1% revenue share for the integrating company.

Can purchases complete without a human approving each one?

Yes, within limits the cardholder set. Purchases inside cardholder-chosen merchant and amount caps, or covered by a revocable app permission with matching checkout conditions, can proceed without a second approval link. Everything else sends the approval link, and every decision is logged.

Conclusion

Broad checkout coverage is what separates a shopping demo from a shopping product, and it is the problem Agentcard was built to solve. OpenCommerce gives your agent one buy tool across 5,000+ stores, Vault keeps customers' own cards encrypted and under their control, and Issuing supplies spend-capped virtual cards when you fund the spend. Choose the products your integration needs; they run side by side.

Start with the Agentcard docs to map your flow, explore OpenCommerce for your purchasing application, and test the whole pipeline in sandbox before your first live order.

Related Articles