Shopify MCP Manager
Enables AI assistants like Cursor or Claude Desktop to fully manage a Shopify store using the Model Context Protocol, with tools for products, orders, customers, inventory, discounts, and analytics.
README
Shopify MCP Manager
An MCP (Model Context Protocol) server for fully managing a Shopify store through AI assistants like Cursor or Claude Desktop.
Architecture
┌─────────────────┐ stdio ┌──────────────────┐ GraphQL ┌──────────────┐
│ Cursor / Claude │◄──────────────►│ MCP Server │◄────────────►│ Shopify API │
│ (MCP Client) │ │ (Node.js) │ │ Admin API │
└─────────────────┘ └──────────────────┘ └──────────────┘
│
▼
┌──────────────────┐
│ token-store.json │
│ (Access Token) │
└──────────────────┘
OAuth Flow
┌──────────┐ 1. /auth/start ┌──────────────────┐
│ Browser │◄────────────────────►│ OAuth Server │
│ │ │ (localhost:3847) │
└─────┬─────┘ └──────────────────┘
│ ▲
│ 2. Redirect to Shopify │ 5. Save token
▼ │
┌──────────────┐ 3. Authorize ┌──────────────────────────────┐
│ Shopify │───────────────►│ Cloudflare Worker │
│ Grant Screen│ 4. Callback │ (your-oauth-callback. │
└──────────────┘ + Auth Code │ workers.dev) │
│ │ │
│ Redirect to localhost:3847 │
│ /callback │
└──────────────────────────────┘
Prerequisites
- Node.js >= 18
- A Shopify App (created in the Dev Dashboard)
- A Cloudflare Worker for URL redirect (OAuth callback)
Quick Start
1. Install dependencies
npm install
2. Configuration
Copy .env.example to .env and fill in your values:
cp .env.example .env
Required values from the Shopify Dev Dashboard:
- SHOPIFY_API_KEY → Client ID of your app
- SHOPIFY_API_SECRET → Client Secret of your app
- SHOPIFY_DOMAIN → Your myshopify.com domain
- SHOPIFY_REDIRECT_URL → Your Cloudflare Worker callback URL
3. Configure Shopify App
In the Shopify Dev Dashboard under your app:
- Set App URL to your Cloudflare Worker domain (e.g.
https://your-oauth-callback.workers.dev) - Add Allowed redirection URLs:
https://your-oauth-callback.workers.dev/callback
4. Configure Cloudflare Worker
Your Cloudflare Worker must redirect OAuth callbacks from Shopify to your local server.
Example Worker:
export default {
async fetch(request: Request): Promise<Response> {
const url = new URL(request.url);
// OAuth callback from Shopify → Redirect to local server
if (url.pathname === "/callback") {
const localUrl = `http://localhost:3847/callback${url.search}`;
return Response.redirect(localUrl, 302);
}
// App URL request from Shopify → Redirect to local server
if (url.searchParams.has("shop")) {
const localUrl = `http://localhost:3847${url.pathname}${url.search}`;
return Response.redirect(localUrl, 302);
}
return new Response("Shopify OAuth Endpoint active.", { status: 200 });
},
};
5. Run OAuth
npm run oauth
This starts a local server and opens the browser. Log in to Shopify and grant the permissions. The access token is automatically saved to token-store.json.
6. Configure MCP Server in Cursor
Add the following to your Cursor configuration (~/.cursor/mcp.json):
{
"mcpServers": {
"shopify-manager": {
"command": "node",
"args": ["/ABSOLUTE/PATH/TO/src/mcp-server.js"],
"env": {
"SHOPIFY_DOMAIN": "your-store.myshopify.com",
"SHOPIFY_API_VERSION": "2026-01"
}
}
}
}
Available Tools (30+)
Shop Management
| Tool | Description |
|---|---|
shop_info |
Shop basics (name, domain, plan, currency) |
shop_locales |
Configured languages |
shop_policies |
Privacy, terms of service, refund policies |
available_shipping_countries |
Shipping destinations |
Products
| Tool | Description |
|---|---|
list_products |
List products with filtering & pagination |
get_product |
Product details with variants & images |
create_product |
Create a new product |
update_product |
Update a product |
delete_product |
Delete a product |
list_collections |
List collections |
Orders
| Tool | Description |
|---|---|
list_orders |
List orders with filtering |
get_order |
Order details |
update_order |
Update an order |
cancel_order |
Cancel an order |
create_draft_order |
Create a draft order |
Customers
| Tool | Description |
|---|---|
list_customers |
List customers |
get_customer |
Customer details |
create_customer |
Create a new customer |
update_customer |
Update a customer |
delete_customer |
Delete a customer |
Inventory
| Tool | Description |
|---|---|
list_locations |
List locations |
get_inventory_levels |
Show inventory levels |
adjust_inventory |
Adjust inventory |
get_product_inventory |
Full product inventory |
Discounts
| Tool | Description |
|---|---|
list_discount_codes |
List discount codes |
create_basic_discount |
Create a discount code |
list_automatic_discounts |
List automatic discounts |
Analytics
| Tool | Description |
|---|---|
get_order_count |
Count orders |
get_product_count |
Count products |
get_customer_count |
Count customers |
shop_dashboard_summary |
Dashboard overview |
graphql_query |
Run arbitrary GraphQL queries |
OAuth Flow Details
The OAuth 2.0 Authorization Code Grant flow works as follows:
- Initiation: User starts the OAuth flow via
http://localhost:3847/auth/start - Redirect to Shopify: The app redirects the user to Shopify's authorization page with:
client_id(App API Key)scope(requested permissions)redirect_uri(Cloudflare Worker URL)state(nonce for CSRF protection)
- User authorizes: The store owner grants the requested permissions
- Callback: Shopify sends an authorization code to the redirect_uri
- Cloudflare Worker: Redirects the callback to the local OAuth server
- Security checks:
- HMAC verification (message authenticity from Shopify)
- State/nonce comparison (CSRF protection)
- Shop domain validation
- Token exchange: The authorization code is exchanged for an access token via
POST https://{shop}/admin/oauth/access_token - Persistence: The token is saved to
token-store.json
Security Notes
- Never commit
token-store.jsonor.env(both are in.gitignore) - The
SHOPIFY_API_SECRETis confidential - Access tokens have the permissions of the configured scopes
- If you suspect token compromise: rotate the token via the Shopify Dev Dashboard
License
MIT
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。