Shopify MCP Server

Shopify MCP Server

MCP server for managing Shopify stores via the Admin GraphQL API, providing tools to read/update products, orders, customers, inventory, and create discount codes.

Category
访问服务器

README

Shopify MCP Server

An MCP (Model Context Protocol) server that gives Claude access to your Shopify store through the Shopify Admin GraphQL API.

It runs in two modes from the same codebase:

  • Remote (HTTP) — deploy it to any host and add it to claude.ai as a custom connector, so it works from the web and mobile apps everywhere. See Use from claude.ai.
  • Local (stdio) — run it on your machine for Claude Desktop or Claude Code.

Tools

Tool What it does
get-shop-info Store name, domain, currency, and plan
get-products List/search products
get-product-by-id Full product details (description, images, variants)
update-product Change a product's title, description, or status
update-variant-price Change a variant's price / compare-at price
get-orders List/search recent orders
get-order-by-id Full order details (line items, address, totals)
get-customers List/search customers
get-customer-orders A customer's profile and recent orders
get-locations List inventory locations
adjust-inventory Add or remove available stock for a variant
create-discount-code Create a percentage or fixed-amount discount code

Read tools accept Shopify search query syntax (e.g. status:active, fulfillment_status:unfulfilled, created_at:>2026-01-01) and IDs may be given as bare numbers or full gid://shopify/... IDs.

Get Shopify API credentials

Both modes need Admin API credentials. There are two kinds, depending on how your app was created; the server supports both.

Dev Dashboard app (current Shopify flow)

Shopify has deprecated creating custom apps inside the store admin — new apps are created in the Dev Dashboard and authenticate with a client ID/secret. The server exchanges these for access tokens automatically (client credentials grant) and refreshes them before their 24-hour expiry.

  1. In the Dev Dashboard, create an app (the "API-only, no admin UI" path is a good fit).

  2. Configure the Admin API access scopes you want the server to have:

    • read_products, write_products
    • read_orders
    • read_customers
    • read_inventory, write_inventory
    • read_discounts, write_discounts
    • read_locations

    (Only grant write_* scopes if you want Claude to be able to make changes. Scopes lock at install time — changing them later means releasing a new version and reinstalling.)

  3. From the app's Home panel, use Install app to install it on your store.

  4. In the app's Settings, copy the Client ID and Client secret. These become the SHOPIFY_CLIENT_ID and SHOPIFY_CLIENT_SECRET environment variables.

Note: the client credentials grant only works for apps created by your own organization and installed on your own store — which is exactly this setup.

Legacy admin custom app

If your store still has an app under Settings → Apps and sales channels → Develop apps with an Admin API access token (starts with shpat_), you can use that instead: set it as SHOPIFY_ACCESS_TOKEN and skip the client ID/secret.

Use from claude.ai (web + mobile)

To always have access without a desktop machine, deploy the server somewhere public and connect it as a custom connector.

1. Deploy the server

Any Node.js host works. The repo includes a Dockerfile, so container platforms like Railway, Render, or Fly.io can deploy it straight from GitHub with no extra config. Set these environment variables on the host:

Variable Value
SHOPIFY_DOMAIN your-store.myshopify.com
SHOPIFY_CLIENT_ID Dev Dashboard app client ID (see above)
SHOPIFY_CLIENT_SECRET Dev Dashboard app client secret
MCP_AUTH_TOKEN A long random secret, e.g. from openssl rand -hex 32

(For a legacy admin custom app, set SHOPIFY_ACCESS_TOKEN instead of the client ID/secret.)

The container listens on PORT (default 3000) and serves:

  • /mcp/<MCP_AUTH_TOKEN> — the MCP endpoint for claude.ai
  • /dashboard/<MCP_AUTH_TOKEN> — a live dashboard webpage (see below)
  • /healthz — health check

Dashboard

https://your-app.example.com/dashboard/<MCP_AUTH_TOKEN> is a bookmarkable web dashboard that pulls live data from Shopify on every load: orders and revenue for today and the last 7 days, new customer signups, an orders-per-day chart, and tables of recent orders (with payment/fulfillment status) and recent signups. It auto-refreshes every 2 minutes and supports light/dark mode. The same MCP_AUTH_TOKEN protects it — treat the URL as confidential.

To run it directly instead of via Docker: npm install && npm run build && npm run start:http.

2. Add the connector on claude.ai

  1. Go to claude.ai → Settings → Connectors → Add custom connector (available on Pro, Max, Team, and Enterprise plans).
  2. Name it Shopify and set the URL to:
    https://your-app.example.com/mcp/<your MCP_AUTH_TOKEN>
    
  3. Save. The Shopify tools now appear in web and mobile chats via the search-and-tools menu.

Security: the MCP_AUTH_TOKEN in the URL is the only thing standing between the internet and your store's API, so make it long and random, always use HTTPS (hosts like Railway/Render provide it automatically), and rotate the token if the URL ever leaks. The Shopify token itself stays server-side and is never sent to the browser.

Run locally (Claude Desktop / Claude Code)

1. Build the server

npm install
npm run build

2. Connect it to Claude

Claude Code (CLI):

claude mcp add shopify \
  --env SHOPIFY_DOMAIN=your-store.myshopify.com \
  --env SHOPIFY_CLIENT_ID=xxxxxxxxxxxx \
  --env SHOPIFY_CLIENT_SECRET=xxxxxxxxxxxx \
  -- node /absolute/path/to/Shopify-MCP/build/index.js

Claude Desktop — add to claude_desktop_config.json:

{
  "mcpServers": {
    "shopify": {
      "command": "node",
      "args": ["/absolute/path/to/Shopify-MCP/build/index.js"],
      "env": {
        "SHOPIFY_DOMAIN": "your-store.myshopify.com",
        "SHOPIFY_CLIENT_ID": "xxxxxxxxxxxx",
        "SHOPIFY_CLIENT_SECRET": "xxxxxxxxxxxx"
      }
    }
  }
}

Then ask Claude things like "What are my 5 most recent unfulfilled orders?" or "Create a 15% discount code called WELCOME15".

Security notes

  • The access token grants API access to your store — treat it like a password. Keep it in host environment variables or your MCP client config; never commit it.
  • Scope the token minimally: if you only need reporting, grant read-only scopes.
  • The server only ever calls your store's /admin/api GraphQL endpoint; the Shopify token never leaves the server.
  • In remote mode, always set MCP_AUTH_TOKEN — without it the endpoint is open to anyone who finds the URL.

Development

npm run watch   # recompile on change

The server is plain TypeScript using @modelcontextprotocol/sdk. Each tool group lives in src/tools/, src/shopify.ts holds the GraphQL client (API version 2025-07), and src/server.ts assembles the MCP server used by both entry points (src/index.ts for stdio, src/http.ts for HTTP).

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选