kitcommerce-mcp
A standalone MCP server that exposes the admin side of kitcommerce-api as tools, enabling AI assistants to manage products, categories, collections, coupons, orders, inventory, customers, and the dashboard.
README
kitcommerce-mcp
A standalone Model Context Protocol server that
exposes the admin side of kitcommerce-api as MCP tools, so an AI assistant
(Claude Desktop, Claude Code, Cursor, …) can manage a store: products,
categories, collections, coupons, orders, inventory, customers and dashboard.
It talks to the API over HTTP — it does not import the API code — so it can point at any deployment (local, staging, prod) via environment variables.
How it works
- Transport: stdio (spawned by the MCP client).
- Auth: sends the admin API key as the
authorizationheader (matched againstApiKey.secretin the API) and the store id asx-litekart-store. - Tools are a curated CRUD set over the core commerce entities (see below).
Setup
cd mcp
npm install
npm run build
Configure credentials (copy .env.example for reference — the MCP client
supplies these as env, see below):
| Variable | Description |
|---|---|
KITCOMMERCE_API_URL |
Base URL of the API, e.g. https://api.litekart.in |
KITCOMMERCE_API_KEY |
Admin API key secret (sent as authorization) |
KITCOMMERCE_STORE_ID |
Store id (sent as x-litekart-store) |
Create the API key in the admin API keys section; it is tied to a store and a user/role, which determines what the MCP server is allowed to do.
Register with an MCP client
Claude Desktop / Claude Code (claude_desktop_config.json or .mcp.json)
{
"mcpServers": {
"kitcommerce": {
"command": "node",
"args": ["D:/projects/lk/kitcommerce-api/mcp/dist/index.js"],
"env": {
"KITCOMMERCE_API_URL": "http://localhost:3000",
"KITCOMMERCE_API_KEY": "your_admin_api_key_secret",
"KITCOMMERCE_STORE_ID": "your_store_id"
}
}
}
}
For local development you can skip the build and run from source with tsx:
{
"command": "npx",
"args": ["tsx", "D:/projects/lk/kitcommerce-api/mcp/src/index.ts"]
}
Tools
| Group | Tools |
|---|---|
| Products | list_products, get_product, create_product, update_product, delete_product, bulk_update_product_status |
| Categories | list_categories, get_category, create_category, update_category, delete_category |
| Collections | list_collections, get_collection, create_collection, update_collection, delete_collection |
| Coupons | list_coupons, get_coupon, create_coupon, update_coupon, delete_coupon |
| Orders | list_orders, get_order, update_order |
| Inventory | list_inventory, get_inventory, update_inventory |
| Customers | list_customers, get_customer |
| Dashboard | get_dashboard_summary |
list_* tools accept page, page_size, search, sort, status, and a
free-form filters object for any additional query params. create_* /
update_* accept a data object sent as the JSON body — call the matching
get_* on an existing record first to learn the exact field shape.
Extending
Tools are declared in src/tools.ts. Add a resource with the crud() helper or
push a bespoke ToolDef. Keep the set curated — most MCP clients cap the number
of tools they will load.
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。