Universal Commerce Protocol (UCP)

UCP is an open standard by Google and Shopify that enables AI agents to handle the full commerce journey from discovery to post-purchase.

Last updated: 2026-10-03

What Is the Universal Commerce Protocol (UCP)?

The Universal Commerce Protocol is an open standard developed by Google and Shopify that lets AI agents handle the full commerce journey, from discovery through checkout to post-purchase, on any compatible platform. It is one of the core protocols behind agentic commerce.

Google announced UCP at the National Retail Federation conference on January 11, 2026. It is the broadest agentic commerce protocol so far, covering the whole lifecycle rather than checkout alone: discovery, browsing, cart, payment, order tracking, returns and customer service.

The protocol defines interactions between four actors: platforms (AI agents and applications), merchants, credential providers (identity) and payment service providers. One design principle matters most to retailers: the merchant stays merchant of record. You keep the customer relationship, the data and the post-purchase experience (Google Developers, January 2026).

UCP was co-developed with Shopify and is endorsed by over 20 major partners including Stripe, Visa, Mastercard, Adyen, American Express, Best Buy, Macy's, Home Depot, Target, Walmart, Wayfair, Etsy, and Zalando (Google Blog, Jan 2026).

Why UCP Matters for Retailers

UCP standardizes the whole journey and connects to Google's Shopping Graph of 50 billion listings.

UCP matters because it standardizes the whole shopping journey rather than one piece of it. Where ACP handles checkout and MCP handles data connectivity, UCP is designed to be the end-to-end framework.

Google's Shopping Graph indexes more than 50 billion product listings, with 2 billion updated every hour (Google, January 2026). UCP is the standard interface agents use to work with that index and complete transactions for shoppers.

It is built to interoperate. UCP works with the Agent Payments Protocol (AP2) for payment, Agent-to-Agent (A2A) for coordination between agents, and MCP for structured data access. Merchants can integrate through REST APIs, A2A or MCP, whichever fits their stack (Google Developers, January 2026).

Early participants in Google AI Mode agentic checkout include Wayfair, Chewy, Etsy and Quince (Google, February 2026). For most other merchants the practical path is through their commerce platform, with Shopify leading.

How UCP Works

Merchants publish a profile and implement the capabilities their systems support: browsing, checkout, post-purchase.

UCP is modular and extensible. Merchants implement capabilities based on what their systems support:

Discovery. Merchants publish a UCP profile at /.well-known/ucp. It advertises services, endpoints, versions and capabilities. It is not the catalog. Google uses Merchant Center data for product discovery, so the feed still does that job.

Browsing and selection. Agents query the catalog, filter by the shopper's preferences and present options. UCP carries rich attributes: images, variants, availability, price tiers, shipping options.

Cart and checkout. The agent manages the cart and starts checkout. Payment runs through AP2-compatible providers and the shopper's credentials are never exposed to the agent.

Post-purchase. Order tracking, returns and customer service routing. This is what separates UCP from checkout-only protocols.

Shopify Engineering described the co-development process: "We co-developed UCP with Google to create an open standard for AI agents to connect and transact with any merchant" (Shopify Engineering, Jan 2026).

UCP vs ACP vs MCP

UCP covers the full lifecycle, ACP focuses on checkout, and MCP is the data connection underneath both.

Scroll horizontally to read the full table →

AttributeUCPACPMCP
DeveloperGoogle / ShopifyStripe / OpenAIAnthropic → Linux Foundation
ScopeFull commerce lifecycleCheckout and paymentsData connectivity
Discovery/.well-known/ucp endpointCommerce feed specificationServer/client connections
PaymentVia AP2 integrationVia StripeN/A (data layer only)
InteroperabilityCompatible with AP2, A2A, MCPStandaloneUniversal (used by UCP, ACP)
Partners20+ (Visa, Mastercard, Best Buy, etc.)Confirm participating merchants with the platform97M+ SDK downloads

The three are complementary. A retailer might use MCP to make catalog data reachable, ACP for ChatGPT checkout and UCP for the Google journey. Whichever protocols your platform supports, they all read the same product information, and that is the part a retailer controls directly. Paz handles that work: a Paz expert keeps the catalog and Merchant Center feed complete and accurate so products are eligible and well described wherever the protocol takes them. See how the product feeds service works. For an engineering walkthrough, see Paz.ai's UCP technical guide for retailers.

FAQ

Is UCP only for Google's AI products?+
No. UCP is an open standard for any AI agent or platform. Google developed it with Shopify, and it is open source and intended for cross-platform use.
Do I need UCP if I already support ACP?+
They serve different platforms. ACP is the ChatGPT path and UCP is the Google path, and each has its own onboarding. For most merchants both arrive through the commerce platform rather than as separate builds. The product information behind them is shared.
How do I implement UCP?+
Follow the versioned specification and Google's onboarding requirements. Implementation starts with a /.well-known/ucp profile that advertises services, endpoints and capabilities. It is not a product feed. Product discovery still runs on Merchant Center data, so the feed and the profile are separate pieces of work.
Does UCP replace my existing product feeds?+
No. UCP works alongside Merchant Center and platform feeds and adds a standard interface for agents on top. Your feeds still carry the product data, which is why their quality still decides what gets found.

Primary sources

Related terms

How AI-ready are your products?

Check one product page for the information gaps that keep it out of AI answers and search results. If you want the fixes handled, Paz experts do that work for ecommerce teams.