Shopify MCP: What It Is and How to Set It Up

Shopify shipped four official MCP servers during Q1 2026, and every eligible store has had a Storefront MCP endpoint live by default since then. Most merchants don't know it's already running. Here's what each server does and how to actually use it.

---

What Shopify MCP is, in one paragraph

Model Context Protocol (MCP) is the open standard that lets AI agents — ChatGPT, Claude, custom shopping assistants — connect to external tools and data sources in a consistent way. Shopify's MCP servers expose a store's catalog, checkout, and customer-account functions to any MCP-compatible AI agent, so an agent can search products, check pricing, read policies, and initiate checkout without a custom integration for every AI platform.

---

The four official servers

Server	Purpose	Status
Storefront MCP	Customer-facing AI shopping agents — product search, cart, pricing, policies	Live by default on every eligible store since Q1 2026
Customer Account MCP	Post-purchase customer interactions (order status, returns, account management)	Live
Checkout MCP	Agentic checkout flows for native applications	Preview, select partners only
Dev MCP	Exposes Shopify's docs, GraphQL schema, and API examples to AI coding assistants	Local-only, for developers building on Shopify

---

Storefront MCP: what it actually does for your store

This is the one that matters most for discovery. It's exposed automatically at your store's standard domain — no opt-in required for eligible stores — and it lets any MCP-compatible AI agent:

• Search your product catalog
• Look up cart and pricing details
• Read your store's policies and FAQs
• Initiate a checkout via the UCP (Universal Commerce Protocol) layer, where scopes allow

Because it's on by default, the practical question for most merchants isn't "how do I turn this on" — it's "is my catalog data actually good enough for an AI agent to use it well." A live MCP endpoint pointed at thin, inconsistent product data doesn't help you; it just means an AI agent can technically query bad data faster.

---

Dev MCP: for your engineering team, not your storefront

This one is easy to confuse with Storefront MCP but serves a completely different purpose. Dev MCP is a local-only server you run on your own machine while building Shopify apps or integrations — it exposes documentation, the GraphQL schema, and sample API calls to your AI coding assistant (Cursor, Claude Code, Continue.dev, and similar tools). It has nothing to do with customer-facing AI shopping agents; it's a developer productivity tool.

---

Setting it up

For Storefront and Customer Account MCP, there's no setup required on eligible stores — they're live by default. Verify your endpoint is active and review documentation at shopify.dev.

For Dev MCP, setup is IDE-specific. In Cursor, for example: go to Cursor > Settings > Cursor Settings > Tools and MCP > New MCP server, add the Shopify Dev MCP configuration, save, and restart. Requirements include Node.js 18 or higher. Documentation and configuration examples for all four servers live at shopify.dev.

For Checkout MCP, access is currently limited to select partners building native applications — check with your Shopify partner rep if you're building a custom agentic checkout flow.

---

Why a live MCP endpoint isn't the same as being AI-visible

Here's the gap most merchants miss: having Storefront MCP live means an AI agent can query your catalog if it decides to. It doesn't mean ChatGPT, Gemini, or Perplexity will actually recommend your products when a relevant shopper asks. Discovery — being the product an AI surface cites or recommends — is a separate problem from technical accessibility, and it depends on catalog data quality, structured schema, and how your brand is discussed across the sources those AI systems draw on.

MCP gives the agent a door in. What's on the other side of that door — how complete, structured, and citable your product data actually is — is what Citation Rank measures and what we help brands improve. If you've confirmed your Storefront MCP is live and still aren't showing up in AI shopping answers, that's the gap to close next. Book a demo for a read on where your catalog stands.

---