Agentspay

Explainer

Universal Commerce Protocol (UCP) Explained: Google and Shopify's Agentic Commerce Standard vs ACP

Google and Shopify shipped UCP as an open standard so an AI agent can check out at any merchant that supports it. Here is what the specification actually defines, where it is live for US buyers, and the one thing it deliberately leaves to you.

Agent Payments Console

Pick an agent

Payment intent

intent:

Policy evaluation

Human approval required

This spend is over your approval threshold. Approve it to issue a scoped card, or deny it.

Scoped virtual card issued

Agentspay

single-use

Wallet budget

spent of

Audit trail

In short

The Universal Commerce Protocol (UCP) is an open standard co-developed by Google and Shopify, announced by Sundar Pichai at the NRF 2026 keynote on January 11, 2026, that lets an AI agent search a catalog, build a cart, and complete checkout at any participating merchant. Businesses publish a machine-readable capability profile at /.well-known/ucp, and agents call the capabilities they find over REST, MCP, A2A, or an embedded binding. The specification is published on GitHub under the Apache 2.0 license and is designed to interoperate with AP2 for payment authorization, A2A for agent-to-agent messaging, and MCP for tool calls. UCP standardizes how the purchase happens. It does not decide whether your agent should have been allowed to make it.

What the Universal Commerce Protocol is

UCP is a common language for commerce, written so that software agents can use it without a human in the loop. The problem it solves is combinatorial: every AI surface that wants to sell things has to understand every merchant's cart rules, discount stacking, tax logic, shipping options, and payment methods, and every merchant would otherwise build a separate integration for every surface. UCP replaces that with one standard vocabulary. A business implements a set of named capabilities, declares which ones it supports, and any agent that speaks UCP can transact with it. Shopify's engineering team described the design as borrowing from TCP/IP: core shopping primitives at the bottom (checkout sessions, line items, totals), named capabilities in the middle, and reverse-domain-named extensions on top that anyone can define without a committee vote. The capability names in the specification look like dev.ucp.shopping.checkout, dev.ucp.shopping.cart, dev.ucp.shopping.catalog.search, dev.ucp.shopping.catalog.lookup, dev.ucp.shopping.order, and dev.ucp.common.identity_linking. That naming scheme is the point: a vendor who owns the domain owns the namespace, so extensions can grow sideways instead of forcing a version bump on the core.

Who developed the Universal Commerce Protocol and who signed on

Google and Shopify co-developed it. Sundar Pichai announced UCP in the NRF 2026 keynote on January 11, 2026, and Google's developer blog and Shopify's engineering blog both published on the same day. The launch coalition Google named runs past twenty partners and mixes retailers, networks, and processors: Etsy, Wayfair, Target, Walmart, Best Buy, Macy's, The Home Depot, Flipkart, Zalando, plus Adyen, American Express, Mastercard, Stripe, and Visa. The presence of the card networks matters more than the retailer names. Visa and Mastercard already run their own agent programs, which we cover in the Visa Intelligent Commerce and Mastercard Agent Pay comparison, and their participation in UCP signals that the networks intend to be the payment handlers inside it rather than compete with it. Governance is genuinely open rather than nominal: the specification lives at github.com/Universal-Commerce-Protocol/ucp under Apache 2.0, with maintainers listed in the repository and contributions accepted through ordinary pull requests.

How the Universal Commerce Protocol works

Discovery comes first. A business publishes a JSON profile at /.well-known/ucp that declares which services and capabilities it supports, which versions, and which payment handlers it accepts. An agent advertises its own profile URL in a UCP-Agent request header. Both sides then compute the intersection of what they mutually support and select the highest shared version of each capability, per transaction. Nothing is hard-coded, which is what lets a merchant add fulfillment options or a new payment handler without every agent re-integrating. Transport is deliberately not fixed. The same capability can be exposed as a REST endpoint described by OpenAPI 3.1.0, as an MCP tool for agents that already speak the Model Context Protocol, over A2A for agent-to-agent flows, or through an embedded binding. Checkout itself is a state machine rather than a single call, and the specification allows escalation: when an agent hits something it cannot handle, the merchant hands back a continue_url and the human finishes in an embedded checkout. That escalation path is the honest part of the design. It admits that autonomous checkout will not cover every cart. UCP orchestrates the purchase; the narrower job of proving the buyer authorized it belongs to Google AP2, which is specified as an extension for UCP rather than a competitor to it.

The four roles UCP defines, and why they matter for payments

The specification names four participants, and the split is where the payments story lives. A Platform is anything that consumes capabilities: an AI agent, a mobile app, a procurement system, or another business. A Business exposes capabilities and is typically the merchant of record. A Credential Provider holds the sensitive user data, meaning payment instruments and shipping addresses. A Payment Service Provider processes the money for the business. Credentials move in one direction only. The platform obtains a payment instrument from the credential provider, passes a token to the business, and the business must not echo credentials back in its responses. Three separate bilateral trust relationships, no shared secret sitting in the agent's context window. That design decision is why prompt injection against a UCP agent cannot exfiltrate a raw card number, though it can still convince the agent to buy the wrong thing, which is a different failure we walk through in prompt injection and agent spend.

UCP vs ACP: what is the difference

The two standards were built by rival coalitions for overlapping jobs, and most US retailers will end up supporting both. The Agentic Commerce Protocol came out of OpenAI and Stripe and powers Instant Checkout inside ChatGPT: the merchant shares a product feed with the platform, the shopper buys without leaving the chat, and the merchant receives a tokenized payment. UCP is merchant-hosted and pull-based. Agents query your live catalog through your own endpoints, so inventory and pricing are current by construction rather than as current as your last feed push. The trade is effort: ACP is quicker to adopt because the platform does the heavy lifting, UCP demands real API infrastructure but gives you control and freshness. Scope differs too. ACP concentrates on the checkout handoff. UCP covers catalog search, cart, checkout, order lifecycle webhooks, and account linking, which makes it closer to a full commerce API than a checkout bridge. The practical answer for a US merchant in 2026 is that these are distribution channels, not architectures you pick between. ACP reaches ChatGPT. UCP reaches AI Mode in Google Search, the Gemini apps, and any open-standard agent. We covered the readiness work either one requires in the agentic commerce readiness guide for merchants, and the fee and setup detail specific to ChatGPT in the Instant Checkout guide.

Is UCP the same as AP2, and where does x402 fit

No, and the confusion is understandable because they ship from the same company. AP2, Google's Agent Payments Protocol, is a payment authorization layer: it carries signed intent and cart mandates that prove a human authorized a purchase within stated limits. UCP is the commerce layer above it, and Google documents the two as interoperable, with AP2 providing the secure payment support inside a UCP flow. Neither is a budget. A mandate proves consent for a purchase; it does not stop an agent from making forty consented purchases in an hour. x402 sits somewhere else entirely, reviving HTTP 402 so a machine can pay a fraction of a cent for a single API response, settled in stablecoins. It is the right tool when an agent buys compute or data per call and the wrong tool when it buys a laptop from Best Buy. If you want the three payment standards side by side rather than the commerce layer, we break them down in AP2 vs ACP vs x402.

What a merchant has to do to accept UCP-powered checkout on Google

Google's Merchant Center documentation is specific about the setup, and it is a short list. Meet the technical requirements in the developer documentation, submit the interest form, add the native_commerce(checkout_eligibility) attribute to your product data, make sure your Merchant Center feed is complete and current, and create a Google Pay and Wallet Console account. We walk through each requirement, the eligibility attribute and the fee question in our UCP checkout setup guide for Google AI Mode and Gemini. Availability at the time of writing is early access, limited to products eligible in the United States, Canada, and Australia, and to participating merchants and partners. It surfaces as a Buy button on product listings inside AI Mode in Search and the Gemini app, paid with credentials the shopper already has in Google Wallet. You stay the seller of record, and Google states merchants may customize the integration to preserve their own checkout requirements. Shopify merchants have a shorter path: Shopify's Checkout Kit implements UCP capabilities for you, which is the practical reason a large share of early UCP volume will come through Shopify stores. What nobody has published is a fee. Google's merchant help page discloses no take rate for UCP-powered checkout, and the protocol itself is free and Apache licensed, so your costs are ordinary payment processing plus whatever engineering the integration takes. Treat any specific UCP commission number you read elsewhere as unverified until Google publishes one. Shopify merchants get the UCP side for free and largely pre-built: every Shopify storefront already publishes a live UCP capability document, which we probed across fourteen US brands in the Shopify agentic commerce guide.

What UCP does not do: the spend-control gap on the buying side

Every one of these standards is written from the merchant's side of the counter. UCP tells your store how to accept an agent order. It says nothing about whether the agent placing that order was supposed to spend $4,000 on cloud credits at 3am. Read the capability list again with a finance hat on: there is no per-agent budget, no approval threshold, no merchant allowlist for the buyer, no way to revoke one agent's ability to transact without breaking the others. That is not an oversight, it is a scope boundary, and it is the same boundary in ACP, AP2, and x402. The controls have to live on the buyer side, in whatever issues the credential the agent presents. Card-level caps from an issuer are the first layer, since an issuer declines at authorization and an agent cannot argue with a decline. Above that you need a policy that spans cards and rails, an approval step for anything unusual, and a record that ties every authorization to a specific agent, a specific task, and the human who authorized the agent. If you are choosing a vendor for that layer, the category map is on our AI agent payment platforms comparison.

Where Agentspay fits

Agentspay is rail-neutral and protocol-neutral by design, which in a market with four competing standards is the useful property. Your agent can check out through UCP at a Google surface, through ACP in ChatGPT, or with a plain card at any US merchant, and the same policy applies before any of it authorizes. Each agent gets a wallet and a hard spend limit that is evaluated before authorization rather than reported afterward, human approval fires above the threshold you set, scoped virtual cards lock spend to specific merchants and single uses, and the policy engine writes an immutable record of who asked, what was approved, and what actually settled. When the next protocol arrives, and it will, the control layer does not need rewriting.

Whatever standard moves the money, Agentspay is the rail-neutral control plane that keeps it governed. See how it works and the control surfaces that enforce policy, approvals, and audit on every transaction.

Side by side

UCP compared with the other agentic commerce protocols

How the four standards differ in scope, discovery, and payment handling. Capabilities as documented by each specification in July 2026; the standards are moving quickly, so verify against the primary spec before you build.

How they compare UCP ACP AP2 x402
Who built it Google and Shopify OpenAI and Stripe Google Coinbase, now Linux Foundation
What it covers Catalog, cart, checkout, orders, identity linking Agent checkout with a merchant Payment authorization and mandates Paying for a single HTTP request
How an agent finds you Capability profile at /.well-known/ucp Product feed shared with the platform Not a discovery layer HTTP 402 response from your endpoint
Catalog freshness Pull from your live endpoints As fresh as your last feed push Not applicable Not applicable
Transports REST, MCP, A2A, embedded REST Signed mandates over agent protocols HTTP
Payment handling Token exchange between credential provider and PSP Tokenized card passed to the merchant Signed intent and cart mandates Stablecoin transfer onchain
License Apache 2.0 on GitHub, open contributions Open spec published by OpenAI Google specification Linux Foundation
Live for US buyers AI Mode in Search and Gemini, early access ChatGPT Instant Checkout Partner pilots Machine-to-machine API payments
Limits what the buying agent can spend No No No, a mandate is not a budget No, metered after the call

Frequently asked

Questions people ask about the Universal Commerce Protocol (UCP)

What is the Universal Commerce Protocol?

The Universal Commerce Protocol is an open standard that lets AI agents discover products, build carts, and complete checkout with any merchant that implements it. Businesses declare supported capabilities in a profile at /.well-known/ucp, and agents call those capabilities over REST, MCP, A2A, or an embedded binding. It covers the full journey, not just payment.

Who developed the Universal Commerce Protocol?

Google and Shopify co-developed UCP and announced it at the NRF 2026 keynote on January 11, 2026, alongside more than twenty launch partners including Etsy, Target, Walmart, Wayfair, Best Buy, The Home Depot, Adyen, American Express, Mastercard, Stripe, and Visa. The specification is maintained in public on GitHub and accepts outside contributions.

How does the Universal Commerce Protocol work?

A business publishes a machine-readable capability profile at /.well-known/ucp listing its supported services, versions, and payment handlers. An agent sends its own profile URL in a UCP-Agent header, both sides compute the intersection of what they support, and the agent then calls capabilities such as catalog search, cart, and checkout over the transport the business exposes.

What is the difference between UCP and the Agentic Commerce Protocol?

ACP, from OpenAI and Stripe, is platform-mediated: you share a product feed and the platform runs checkout inside ChatGPT. UCP is merchant-hosted: agents query your live endpoints, so catalog data is current by construction. UCP also spans catalog, cart, orders, and identity linking, while ACP concentrates on the checkout handoff.

Is UCP the same as AP2?

No. AP2 is Google's Agent Payments Protocol, a payment authorization layer that carries signed intent and cart mandates proving a human approved a purchase. UCP is the commerce layer above it and is documented as interoperable with AP2. In practice UCP runs the shopping journey and AP2 can secure the payment step inside it.

Is the Universal Commerce Protocol open source?

Yes. The specification and documentation are published at github.com/Universal-Commerce-Protocol/ucp under the Apache License 2.0, with maintainers listed in the repository and contributions accepted through pull requests. Anyone can implement it without a licensing agreement, and vendors can add extensions in their own reverse-domain namespace.

Does UCP charge merchants a fee?

The protocol itself is free and Apache licensed. Google's Merchant Center documentation for UCP-powered checkout does not publish a take rate, so the disclosed costs are your normal payment processing and the engineering to implement the capabilities. Any specific UCP commission figure circulating in third-party posts is unverified.

Does UCP limit how much an AI agent can spend?

No. UCP standardizes how a merchant accepts an agent order and defines nothing about the buyer's budget. There is no per-agent limit, approval threshold, or merchant allowlist for the purchasing side. Those controls belong to whatever issues the credential the agent presents, which is the layer a control plane provides.

Keep reading

More explainers

ServiceNow AI Control Tower

ServiceNow AI Control Tower

ServiceNow AI Control Tower is the most complete agent inventory and risk console a large US enterprise can buy, and it now reaches across AWS, Google Cloud and Azure. We read the schema ServiceNow ships to developers to answer the one question the rollout meeting always ends on: can it stop an agent from spending money? It cannot, and the reason is written into the data model.

Read

Gemini Enterprise

Gemini Enterprise

Google did something in August 2026 that the other agent platforms have not done: it shipped a hard monthly spend cap that genuinely stops usage instead of emailing you about it. That deserves credit, and it also moves the interesting question one step along. A cap that stops something is only as useful as the thing it is scoped to, so we went and measured what Google can actually point that cap at, in the API model Google publishes for anyone to read.

Read

Salesforce Agentforce

Salesforce Agentforce

Agentforce is the largest agent platform any US enterprise is likely to already own, and it moved to consumption billing, which means the meter now runs on what your agents do rather than on how many seats you bought. That raises a finance question the rollout deck rarely answers: when an Agentforce agent is loose in production, what actually stops it spending. We went and measured the answer in Salesforce own published object model rather than guessing at it.

Read

AWS AgentCore

AWS AgentCore

Amazon shipped the missing piece in August 2026. Bedrock AgentCore Payments went generally available, and it is a real payments product: an agent can now hold a wallet, meet an HTTP 402, pay, and carry on reasoning without a human in the loop. So the question a platform lead has to answer stopped being whether AWS gives agents money and became a narrower, more awkward one: how much of a spend policy did AWS actually ship? We went and measured it, property by property, in the API model AWS publishes.

Read

Microsoft Agent 365

Microsoft Agent 365

Microsoft shipped a control plane for AI agents, and it is a good one. It gives every agent an identity, a registry entry, an owner, a sponsor and a Conditional Access policy. Then somebody in finance asks the obvious follow-up question: fine, but what stops the agent from spending money? This page answers what Agent 365 costs, what it governs, and what we measured when we went looking for a dollar amount anywhere in Microsoft's agent governance surface.

Read

QuickBooks MCP Server

QuickBooks MCP server

Connecting an accounting system to an AI assistant is now a ten minute job. Deciding what that assistant is allowed to do once it is connected is the part nobody writes about, and it is the part your controller will ask about first. This page compares what the official QuickBooks, NetSuite and Xero MCP servers actually hand a model, measured rather than summarized from marketing pages.

Read

Payment MCP Servers

payment MCP servers

Every large payment company shipped an MCP server in the last eighteen months, and almost every write-up of them is a setup tutorial. The setup is the easy part. The question worth answering before you connect one to a production account is narrower and much less comfortable: what, exactly, can the model on the other end of that connection do to your money?

Read

PayPal Agentic Commerce

PayPal Agentic Commerce

PayPal made a bet that most merchants would rather not implement a commerce protocol at all. Where Stripe and OpenAI shipped a spec for you to build against, PayPal shipped two products that sit on top of the checkout you already have, and then bought a company to make the catalog half work. That choice is the whole story: it explains why Agent Ready needs almost no engineering from you, why there is nothing for an agent to discover about your store on the open web, and why the thing PayPal will not do for you is the thing that gets expensive later.

Read

Shopify Agentic Commerce

Shopify Agentic Commerce

Shopify switched agentic commerce on by default, so your store is probably already selling to AI assistants whether or not anyone on your team configured it. Instead of restating the announcement, we checked something you can check too: on September 2, 2026 we requested the machine-readable capability file that Shopify publishes for real storefronts, on fourteen well-known US brand domains, and read what it exposes to an agent. Eleven answered correctly. The three that did not share one trait, and it is quietly costing them agent traffic.

Read

Web Bot Auth

Web Bot Auth

Web Bot Auth is the reason your agent either gets served or gets throttled with the scrapers. Almost everything written about it repeats the same architecture diagram, so we did something different: on September 1, 2026 we fetched the published key directories of more than twenty major AI operators and infrastructure vendors to see who is genuinely signing their traffic. Four were. The results are in the first table.

Read

Tempo Blockchain

the Tempo blockchain

Tempo is the payments chain Stripe and Paradigm built, and it shipped with a protocol that lets software pay for things on its own. It settles machine payments in under a second. It has nothing at all to say about whether your agent should have paid.

Read

AI Agent Governance

AI agent governance

Every agentic AI governance framework published so far governs the same four things: identity, tools, data and prompts. Not one of them carries a budget. Here is what the real frameworks say, which guardrails actually bind at runtime, and what to do about the last mile none of them reach.

Read

A2A Protocol

A2A Protocol

Most explanations of the A2A protocol stop at the sentence that agents can now talk to each other. That was true in April 2025 and it is no longer the interesting part. A2A shipped version 1.0 in April 2026 under Linux Foundation governance, it runs in production inside Azure AI Foundry and Amazon Bedrock AgentCore, and the questions engineers actually get stuck on are narrower: what an Agent Card commits you to, when to reach for MCP instead, and what happens the first time one of your agents has to pay another one for the work. That last question has a specific answer, and it is not in the core spec.

Read

Mastercard Agent Pay

Mastercard Agent Pay

Nearly every article about Mastercard Agent Pay is a retelling of one press release from April 2025, the one where Mastercard said AI agents would be able to shop with Agentic Tokens and named Microsoft as the first platform. That was sixteen months ago, and four more things have shipped since. Reading only the launch coverage leaves you with roughly a quarter of the picture, and the missing three quarters are the parts that decide whether you can actually put this into production.

Read

Visa Intelligent Commerce

Visa Intelligent Commerce

Almost everything written about Visa Intelligent Commerce is a retelling of the April 2025 announcement, when Visa said AI agents would be able to pay with a Visa credential. Three more things have shipped since, including an open agent-identity protocol built with Cloudflare that most coverage does not mention at all. This page is the current version, checked against Visa’s own developer documentation and newsroom in August 2026.

Read

Stripe agentic commerce

Stripe agentic commerce

Most writing about Stripe and agentic commerce is still a retelling of the September 2025 launch week, when Stripe and OpenAI shipped Instant Checkout and published the Agentic Commerce Protocol together. Stripe has built a good deal more since then, and some of it points in a direction the launch coverage never anticipated. This page is the current version, checked against Stripe’s own documentation in August 2026.

Read

ChatGPT Instant Checkout

ChatGPT Instant Checkout

Almost every guide to ChatGPT Instant Checkout still reads like it was written the week it launched, walking merchants through how to apply and what the fee will be. OpenAI changed course in March 2026. Here is the accurate version: what Instant Checkout was, what the numbers actually looked like, what replaced it, and which parts of the stack are still very much alive.

Read

Google AP2

Google AP2

Most guides to Google AP2 still describe an Intent Mandate and a Cart Mandate, because most of them are rewrites of the September 2025 launch post. The specification moved. Here is what the Agent Payments Protocol actually defines today, and the one question it deliberately does not answer.

Read

Human in the loop AI

Human in the Loop AI

Every guide to human in the loop AI describes the same shape: the agent pauses, a person decides, the agent continues. The shape is right. What almost none of them ask is a harder question, which is where the pause is enforced, because a pause written into the agent's own code is a pause the agent is trusted to honor.

Read

AI agent cost

AI Agent Cost

Every cost guide for AI agents answers the same two questions: what does it cost to build, and what does it cost to run. Both are answerable, and both are on somebody's invoice. The third question is the one that ends up in a variance report, because the agent also spends your money, and nobody sends you a bill for that.

Read

Agentic checkout

Agentic Checkout

Nearly every guide to agentic checkout is written for the merchant who wants to receive these orders. Far fewer are written for the company whose agents are placing them, which is odd, because agentic checkout quietly removes the one screen where spending used to get a second look.

Read

API monetization

API Monetization

Most guides to API monetization argue about which pricing model wins. The harder question in 2026 is who is calling. An API priced for a signed-up developer with a key behaves very differently when the caller is an agent that showed up once, wants one record, and has no account.

Read

x402 protocol

x402 Protocol

x402 took the one HTTP status code the web never used and turned it into a payment rail machines can drive. The protocol is elegant and genuinely small. The part it deliberately leaves to you is the budget.

Read

AI procurement agents

AI Procurement Agents

Every major procurement suite shipped agents during 2026. Almost none of them answer the question your controller will ask first, which is what happens when the agent is wrong about a purchase and the money has already moved.

Read

Agentic payments

Agentic Payments

Agentic payments move money with no human at the checkout. The rails to do it all shipped during 2026. The part most teams have not solved is deciding, before the money moves, whether the agent was allowed to spend it.

Read

AI agent monetization

AI Agent Monetization

Every AI agent company is rewriting its price list. The models that survive are metered. The ones that quietly fail are the ones where nobody measured what a single task costs to serve.

Read

Agent payment platforms

AI Agent Payment Platforms

Five different kinds of product now call themselves an AI agent payment platform, and they solve five different problems. Picking the wrong category is the expensive mistake, not picking the wrong vendor inside a category.

Read

MCP Payments

MCP Payments

MCP payments are how an AI agent discovers a payment tool and calls it to move money. The catch: the Model Context Protocol carries the tool call, not the spending decision, so nothing in the stack asks whether the purchase should have happened.

Read

Visa Intelligent Commerce vs Mastercard Agent Pay

Visa Intelligent Commerce vs Mastercard Agent Pay

Visa Intelligent Commerce and Mastercard Agent Pay are the two big card networks racing to let AI agents pay. They take different routes to the same idea, and neither one decides whether a given purchase should have happened.

Read

Agentic Commerce Protocol

the Agentic Commerce Protocol

ACP is the open standard behind agentic checkout in ChatGPT. It tells a merchant how to sell to an AI agent. It says nothing about whether your agent should have made the purchase.

Read

AP2 vs ACP vs x402

AP2 vs ACP vs x402

AP2, ACP, and x402 are the three standards shaping how AI agents pay. They solve different layers of the problem, and most real systems will touch more than one.

Read

Machine payments protocol

Machine payments protocol

As software starts paying software, machine payments protocols define how value moves without a human at the keyboard. The harder question is how to keep that spending governed.

Read

Know Your Agent (KYA)

Know Your Agent

KYA, or Know Your Agent, extends the idea of customer due diligence to autonomous software. When an agent spends, you need to know which agent, on whose authority, and under what limits.

Read

Keep agent spending governed

Add policy, hard limits, human approval, and an immutable audit trail across any protocol or rail. Start in the sandbox today.

Never moves money without policy