Shopify MCP Server

Shopify MCP Server

Shopify MCP Server is an implementation of the Model Context Protocol specifically designed to enable integration between AI applications and various Shopify APIs, making it easy to automate, query, and utilize Shopify developer resources and documentation via standardized AI tooling.

Author: Shopify


View Protocol

What is Shopify MCP Server?

Shopify MCP Server is a lightweight server that exposes a set of tools and prompts to language model-powered applications, allowing them to interact directly with Shopify’s developer APIs and documentation. By running this server, you grant AI agents seamless, protocol-standardized access to key Shopify platform functionalities, such as searching documentation, exploring GraphQL API schemas, and retrieving onboarding information.

How to Configure

To set up the Shopify MCP Server, you just need Node.js and npx installed. Use the following command to start the server:

npx -y @shopify/dev-mcp@latest

To integrate with AI applications like Cursor or Claude Desktop, add the following configuration:

{
  "mcpServers": {
    "shopify-dev-mcp": {
      "command": "npx",
      "args": ["-y", "@shopify/dev-mcp@latest"]
    }
  }
}

On Windows, you may use:

{
  "mcpServers": {
    "shopify-dev-mcp": {
      "command": "cmd",
      "args": ["/k", "npx", "-y", "@shopify/dev-mcp@latest"]
    }
  }
}

To disable telemetry, set the OPT_OUT_INSTRUMENTATION environment variable:

{
  "mcpServers": {
    "shopify-dev-mcp": {
      "command": "npx",
      "args": ["-y", "@shopify/dev-mcp@latest"],
      "env": {
        "OPT_OUT_INSTRUMENTATION": "true"
      }
    }
  }
}

For experimental Polaris Web Components documentation, use:

{
  "mcpServers": {
    "shopify-dev-mcp": {
      "command": "npx",
      "args": ["-y", "@shopify/dev-mcp@latest"],
      "env": {
        "POLARIS_UNIFIED": "true"
      }
    }
  }
}

How to Use

  1. Launch the Shopify MCP Server using the appropriate npx command or your configured MCP integration.
  2. Connect your LLM application (like Cursor or Claude Desktop) to the server using the provided configuration, allowing direct access to Shopify API tools and documentation.
  3. Through your AI application’s UI, discover available tools or prompts, invoke functions like searching developer documentation, exploring the Admin GraphQL schema, or fetching documentation directly.
  4. Optionally, enable or disable additional features like Polaris documentation via environment variables as needed.

Key Features

  • Connects AI apps with Shopify APIs and resources via a standardized protocol
  • Exposes tools for searching, retrieving documentation, and accessing GraphQL schemas
  • Secure, local hosting with flexible environment variable configuration
  • Supports both basic and advanced developer onboarding workflows
  • Extensible via MCP, making it compatible with various AI/LLM clients
  • Optional support for Polaris Web Components documentation (experimental)

Use Cases

  • Rapidly answer developer queries about Shopify APIs and usage
  • Write, introspect, and debug Admin GraphQL API queries and mutations interactively
  • Automate retrieval and surfacing of up-to-date documentation to developer tools or chatbots
  • Onboard new developers with direct, AI-assisted access to reference material and API guides
  • Integrate Shopify backend capabilities into agent workflows or IDE extensions using LLMs

FAQ

Q1: Can I use Shopify MCP Server with AI tools other than Cursor or Claude Desktop?
A1: Yes. As long as your application supports the MCP protocol, you can connect it to the Shopify MCP Server.

Q2: What permissions does the server need?
A2: The server needs network access to communicate with Shopify’s APIs and read access to any local resources it integrates with. No additional privileges are required unless custom code is added.

Q3: How do I update to the latest version?
A3: Simply re-run the npx command with "@latest" to always obtain and use the most recent release of the package.

Q4: Is any of my data sent to Shopify or third parties?
A4: The server performs limited instrumentation by default for internal improvement, but you can disable all telemetry with the OPT_OUT_INSTRUMENTATION environment variable.

Q5: How do I add my own custom tools or prompts?
A5: You can clone and modify the source using the MCP SDK, then run your custom build as an MCP server. See the development section for build instructions.