The standard plan charges 1% of each completed order, so a $100 order carries a $1 platform fee. Payments run on Agentcard, with Stripe Link and Shop listed as “coming soon”. The site says a warmed route buys in 9 seconds against 2 minutes 40 seconds cold.
Business Pill · DISTRIBUTION ADVANTAGE
A short Business Engineer explainer on distribution advantage: why access to customers and channels can matter as much as the product. It teaches the general idea only and does not discuss Agentcard or opencommerce.
The key insight: Agentcard has put a price on agent checkout: 1% of each completed order through one buy tool across 5,026 stores, launched a day after Mastercard’s chief executive called a virtual card in the middle of agent purchases a bridge.
Developers add a single “buy tool” to an agent, according to the site. The agent searches for products, adds the selected one to the cart, confirms the cart and requests payment approval, then sends purchase updates to the shopper. The site’s sample code runs the tool with a Claude model through the AI SDK.
The announcement, posted by the account @keyserfaty, whose X bio says it is building Agentcard in Y Combinator’s S26 batch, described three problems the product is built to solve: coverage, reliability and speed.
On coverage, it said agentic purchasing “shouldn’t involve browser automations or dealing with PSPs”. On reliability, it claimed “100% uptime” and said the team runs thousands of tests every couple of hours on every one of the stores. On speed, it promised responses in under 10 seconds.
First-response times for the 20 named merchants with product search on the first page of opencommerce.sh’s monitoring table, read 31 October 2026. Company-published figures; median and ratio are FourWeekMBA arithmetic.
What the Monitoring Page Shows
The site publishes a live monitoring table that tests each monitored store from search to checkout. When we read it on 31 October, the 20 named merchants with product search shown on its first page each carried a 100.0% success rate.
Their first-response times ran from 4.4 seconds at Uniqlo to 22.7 seconds at DoorDash, with Amazon at 8.2 and Walmart at 7.4. By our arithmetic the median of those 20 is 10.45 seconds, and 19 of the 20 are under 20 seconds. The site rounds the warmed-versus-cold gap to “18× faster”; by our arithmetic 160 seconds over 9 is 17.8 times.
The table also lists Agentcard’s own flights, hotels, classes, reservations and tickets routes, and eBay, McMaster-Carr and TikTok Shop marked “no product search”. It runs to 91 pages.
The Business Lever: Who Owns Agent Checkout
The pricing page sets the business model: Standard charges 1% of each completed order across every merchant and payment provider, and Enterprise adds API access with custom fees. Agentcard’s homepage says consumer apps, B2B companies and crypto teams run their agents’ purchases on it, naming Orchid, Almanac and Laso Finance.
The launch lands a day after Mastercard’s chief executive, Michael Miebach, told analysts on the company’s third-quarter call that a virtual card number in between is “a bridge to get ready for a fully tokenized transaction”, and “not a long-term solution that we will support”.
In the post that flagged the launch, the X account @gaspardlezin asked: “who owns this checkout once network tokens work everywhere?”
The Structural Read
As we read the launch, the product sells coverage to agent builders: one integration instead of one per store, with the fee charged only on completed orders.
As we read it, the published monitoring table is part of the pitch: by our arithmetic the 20 named merchants on its first page show a median first response of 10.45 seconds.
As we read it, the open question is the payment rail: Mastercard said on its call that a virtual card number in between is not a long-term solution it will support, and the pages read do not say which rail opencommerce uses.
Announcement of opencommerce on X, 30 Oct
“searching for products and buying on behalf of users shouldnโt be as painful as it is today”
Three Implications
FOR AGENT BUILDERS The site prices the Standard plan at 1% of each completed order, across every merchant and payment provider.
FOR MERCHANTS The site lists Amazon, Walmart, Target and DoorDash among stores it tests from search to checkout, each shown at a 100.0% success rate.
FOR CARD NETWORKS Mastercard’s chief executive said on 29 October that a virtual card number in between is a bridge to a fully tokenized transaction.
As we read it, opencommerce pitches itself the same way: one endpoint that agents call instead of integrating each store. The framework was published in December 2025 and does not discuss Agentcard or opencommerce.
What Is Not Established
The pages read do not say how opencommerce reaches each store, whether its purchases use virtual card numbers or network tokens, or whether Amazon, DoorDash or the other named merchants have agreed to the integration. They do not give order volumes, revenue or funding. The success rates and speeds are the company’s own published figures.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
Agentcard prices agent checkout at 1% of each completed order across 5,026 stores, through one buy tool. It launched a day after Mastercard’s chief executive called a virtual card in the middle of agent purchases a bridge, not a long-term solution.
A note on sourcing. Product facts, prices and performance figures come from opencommerce.sh and the launch post on X, read on 31 October; success rates and speeds are the company’s own. The Mastercard quotes come from the company’s third-quarter call transcript. The lead came from a post on X. We did not contact Agentcard.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
