Universal Commerce Protocol

What is UCP?

Universal Commerce Protocol (UCP) is Shopify's open infrastructure for enabling AI agents to participate in commerce. It provides standardized APIs, data formats, and interaction patterns that allow AI systems to search products, understand availability and pricing, and facilitate purchases across millions of merchants.

Think of UCP as the “roads and signage” for AI shopping. Just as humans need storefronts, search bars, and checkout flows to shop online, AI agents need structured data and APIs to do the same. UCP provides that infrastructure.

The Foundation: Product Taxonomy

Before AI agents can discover your products, they need to understand them. This starts with Shopify's product taxonomy, a standardized system for categorizing products that provides first-party data to enhance discoverability.

Hear From Shopify

We sat down with Shopify's UCP team to understand their vision for agentic commerce. Click any card to watch the interview segment.

Why Does UCP Matter?

The way people shop is changing. Instead of visiting individual stores or browsing marketplaces, users increasingly ask AI assistants for help. “Find me a good laptop for video editing under $1500” or “What's the best gift for a 10-year-old who likes science?”

Without protocols like UCP, AI agents can only scrape websites or use limited search APIs. With UCP, they have direct access to structured product data, real-time inventory, verified pricing, and merchant information, everything needed to make informed recommendations.

For Merchants: Why Participate?

For Developers: Why Build on UCP?

The Three UCP Catalog Tools

The Storefront Catalog MCP exposes three tools for agents to discover and inspect products on a single merchant. Each is reached via JSON-RPC 2.0 at POST https://{store}/api/ucp/mcp.

Endpoint and Request Shape

UCP uses JSON-RPC 2.0. Every call targets the tools/call method on the store's UCP endpoint, identifies the tool in params.name, and passes the tool inputs under params.arguments. An agent profile URL under meta.ucp-agent.profile declares which UCP capabilities the caller supports.

POST https://examplestore.myshopify.com/api/ucp/mcp
Content-Type: application/json

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "search_catalog",
    "arguments": {
      "meta": {
        "ucp-agent": {
          "profile": "https://shopify.dev/ucp/agent-profiles/2026-04-08/valid-with-capabilities.json"
        }
      },
      "catalog": {
        "query": "organic coffee beans",
        "context": {
          "address_country": "US",
          "language": "en",
          "currency": "USD",
          "intent": "browse"
        }
      }
    }
  }
}

Buyer Context

Agents can pass buyer signals via catalog.context. UCP uses these to localize pricing, assess shippability, and tune relevance.

Cursor Pagination and Match Correlation

UCP search responses include a cursor for deep pagination rather than offset-based paging. Clients advance by passing the returned cursor back on the next call while has_next_page is true. A total_count accompanies the first page.

The batch lookup_catalog tool returns an inputs[] array correlating each request id with the product it resolved to, along with a match_type:

• "exact": the id resolved to that exact product or variant.
• "featured": the id did not resolve, but the store substituted a featured alternative.

Unresolved ids appear in not_found[] with a reason, so agents can decide whether to surface or silently drop stale identifiers.

UCP and the Catalog API

The Catalog API is the primary interface for product discovery within UCP. It provides federated search across all participating merchants, with filtering for price, category, availability, and more.

Catalog Search (this tool) is a reference implementation showing how to query and display results from the Catalog API. Developers can use it to understand response formats and view example curl commands in the documentation.

Where This Is Headed

Agentic commerce is in its early stages. Today, AI assistants can search and recommend products. Tomorrow, they'll handle increasingly complex purchase flows: comparing options across dimensions, negotiating on behalf of users, and managing subscriptions.

UCP is Shopify's bet on this future. By building open infrastructure now, they're positioning Shopify merchants to benefit as AI shopping becomes mainstream.

• Integration with major AI platforms (OpenAI, Google, Anthropic)
• Expanded checkout capabilities for AI-facilitated purchases
• Richer product data including reviews, Q&A, and merchant reputation
• Tools for merchants to optimize their catalog for AI discovery