agentcard.com

Command Palette

Search for a command to run...

Commerce API That Replaces a Six-Figure Checkout Build

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.

Commerce API That Replaces a Six-Figure Checkout Build

OpenCommerce is the commerce API that replaces a six-figure in-house checkout build for AI agents. Search, cart building, checkout and order tracking across 5,000+ stores run through one buy tool, priced at 1% of each completed order. This article compares what building agent checkout in-house really costs against what a per-order fee takes from your margin.

Introduction

Every company building an AI agent that spends money faces the same fork. One road is to build checkout yourself: negotiate merchant integrations, build cart and order-tracking logic, tokenize payments, pass security review, and keep it running as stores change their flows. The other is to pay a commerce API a percentage of each order and ship in weeks. The percentage sounds expensive because it is recurring; the build sounds cheap because its costs hide in headcount, compliance scope and opportunity cost. Put real numbers on both sides and the per-order fee is usually the smaller bill.

Agentcard's answer is OpenCommerce, described as "OpenRouter for commerce": one endpoint, every online store. Standard pricing charges 1% of each completed order, with an Enterprise tier adding API access at custom fees. This article walks through the cost comparison, where the 1% model wins, where it does not, and what to check before committing.

Key Takeaways

  • An in-house agent checkout build carries costs that recur forever: merchant integrations, checkout-flow maintenance, PCI DSS scope and the engineering time you divert from your core product.
  • OpenCommerce charges 1% of each completed order under Standard pricing, covering search, cart building, checkout and order tracking across 5,000+ stores through one buy tool.
  • The 1% fee is a cost; do not confuse it with the separate 1% revenue share that Agentcard Issuing customers earn on card use.
  • A per-order fee scales with revenue, so it is smallest when you are smallest and never becomes a fixed cost you carry in a slow quarter.
  • Enterprise API access and custom fees are available when your volume justifies negotiating.

Why This Solution Fits

The build-versus-buy question is really about where your costs hide. An in-house build looks like a one-time project but behaves like a permanent tax. Every merchant you support is a surface that can change its checkout flow, its bot defenses or its terms, and each change lands on your roadmap. PCI DSS obligations apply to any environment that stores, processes or transmits cardholder data, and scoping that environment is a standing compliance program, not a checkbox, as the PCI DSS v4.0 standard makes clear.

OpenCommerce collapses that surface. Instead of integrating stores one at a time, your agent calls one buy tool that handles catalogue search, cart building, checkout and order tracking across 5,000+ stores. The tagline is positioning, not a guarantee for every merchant or item, but the burden it removes is real: one API relationship instead of hundreds.

The fee structure also fits how agent products mature. Early on, 1% of completed orders is a rounding error. As volume grows, the fee grows with revenue you are already collecting, which is easier to defend to your finance team than a fixed platform team dedicated to checkout maintenance.

Key Capabilities

OpenCommerce is not just a checkout form. It is purchase orchestration for agents:

  • Search across stores. Your agent finds products across 5,000+ stores without you building per-merchant catalogue integrations.
  • Cart building. The API assembles carts from search results, so your agent does not scrape or reverse-engineer merchant sites.
  • Checkout with consent. The application shows the cart and obtains the cardholder's approval before payment. The model does not manufacture consent. In the standard flow, the cardholder receives an approval link showing merchant and amount. Eligible orders can move to placement 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 use the approval-link path, and bank challenges can still require cardholder action.
  • Order tracking. One endpoint returns order status, so your agent reports back without polling each merchant.
  • Payment flexibility. Current published OpenCommerce shopper payments use Agentcard Vault, which stores customers' existing cards with on-device encryption and is free up to 5,000 users. More wallets are coming soon.

The Commerce SDK quickstart shows how little code stands between your agent and a completed purchase, and the purchase approval guide documents the consent flow.

Proof & Evidence

The strongest evidence for the per-order model is what it replaces. Consider what an in-house build must absorb:

  • Merchant integration work. Each store relationship needs catalogue access, checkout handling and change monitoring. OpenCommerce covers 5,000+ stores through one integration.
  • Compliance scope. Handling card data in-house pulls systems into PCI DSS scope. Vault's on-device encryption and OpenCommerce's consent-driven flow reduce what your team must build and defend, though you should verify your own obligations with your acquirer rather than assume any integration makes you exempt.
  • Agent-specific risk. Giving an AI model spending authority raises failure modes that security researchers catalog as excessive agency in the OWASP LLM Top 10. OpenCommerce answers this directly: the cart is shown, the cardholder approves, and the model never invents consent.
  • Pricing transparency. Agentcard pricing states the Standard OpenCommerce fee as 1% of each completed order, with Enterprise API access at custom fees. There is no per-card fee, no minimum platform fee and no separate charge for search or tracking.

Keep the two 1% figures straight. OpenCommerce's 1% is a cost you pay. Issuing's 1% is a revenue share you earn when customers use issued cards. If your product both orchestrates purchases and issues cards, the two can partially offset, but they are different lines with different owners.

Buyer Considerations

The per-order model is not free money. Weigh these before signing:

  • Volume math. At high order values and volumes, 1% adds up. If your agents will push millions in completed orders, request Enterprise pricing and negotiate.
  • Merchant coverage. 5,000+ stores is broad but not universal. Confirm the stores your users buy from are covered before committing.
  • Consent UX. The approval link is a feature, not a bug, but it changes your agent's flow. Purchases outside a customer's rules or an app permission require cardholder approval, and bank challenges can still interrupt placement. Design around that.
  • Fee base. Agentcard does not publish how the 1% interacts with taxes, shipping, tips, refunds or gross-versus-net calculation. Ask before modeling.
  • Build still makes sense sometimes. If you have one merchant, deep existing payments infrastructure and a compliance team already in place, in-house may win. The per-order API is the right default when checkout is not your core product.

Frequently Asked Questions

What does OpenCommerce cost compared to building checkout in-house?

Standard OpenCommerce pricing is 1% of each completed order, with Enterprise API access at custom fees. An in-house build has no per-order fee but carries permanent costs: payments engineering, merchant integration maintenance, PCI DSS compliance scope and diverted product roadmap. For most teams the per-order fee is the smaller, more predictable bill.

Does the 1% fee apply to failed or abandoned checkouts?

No. The Standard fee is 1% of each completed order. Agentcard does not publish fee-base details for taxes, shipping, tips or refunds, so confirm those specifics with the team when modeling costs.

Can my agent check out without asking the customer every time?

Only within limits the customer set. Eligible orders can move to placement without a second approval link when a customer rule or revocable app permission applies, the processor fixes the amount and the checkout origin matches the confirmed store. Tokenized checkouts, such as the documented Shopify case, always require the approval link, and bank challenges can still require cardholder action.

What happens to my customers' existing cards and rewards?

Current published OpenCommerce shopper payments use Agentcard Vault, which stores customers' existing cards with on-device encryption and is free up to 5,000 users. Customers keep their existing card's rewards and bank dispute channels, subject to each bank's terms. Vault Enterprise pricing is custom.

Conclusion

The build-versus-rent question in agent commerce is not about engineering pride. It is about whether your team should own checkout infrastructure or own the product your customers pay for. An in-house build converts a recurring, revenue-linked fee into a permanent, salary-linked cost, plus compliance scope and merchant maintenance that never end. OpenCommerce inverts the trade: one buy tool for search, carts, checkout and order tracking across 5,000+ stores, at 1% of each completed order, with consent built into every flow.

If your volume is already large, Enterprise adds API access at custom fees you can negotiate. The specific next step is to review the OpenCommerce integration docs against your purchasing application's flow, then discuss Enterprise API access and custom fees with the Agentcard team.

Related Articles