Shadcn Registry manager
MCP server for Shadcn UI, enabling automated, remote, or containerized project management via local or remote registries
README
shadcn registry manager - MCP Server
This project provides an MCP (Model Context Protocol) server for the shadcn/ui component registry and CLI. It enables remote, programmatic, or containerized execution of shadcn CLI commands (such as init, add, etc.) that you would normally run locally, making it easy to automate, integrate, or run in cloud/dev environments. The MCP server can connect to both local and remote shadcn registries, providing a secure, extensible, and technical interface for automating project management tasks via AI or CLI agents.
What It Does
This MCP server exposes shadcn CLI operations as MCP tools, so you (or an AI agent) can:
- Initialize a project
- Add components from the shadcn registry including block components
- List and fetch registry items
- Run all shadcn CLI workflows remotely
- Supports both the official shadcn registry - https://ui.shadcn.com/r and custom/self-hosted registries follow this guide: https://ui.shadcn.com/docs/blocks
Usage
Docker
Run the MCP server in a container, mounting your project directory:
Add the following to your mcp.json file:
"shadcn-registry-manager": {
"command": "docker",
"args": [
"run",
"--rm",
"-i",
"--mount",
"type=bind,src=<your-project-path>,dst=/workspace",
"-e",
"REGISTRY_URL=https://ui.shadcn.com/r",
"-e",
"STYLE=new-york",
"reuvenaor/shadcn-registry-manager"
],
},
Docker mounting explanation:
-
<your-project-path> - Is the path to your project directory.
Example:
/Library/Projects/shadcn-registry-manager/www -
/workspace - is the path to the workspace directory inside the container. (Don't change it, if you do, you must pass the same path as
cwdparam on every tool call)
Docker ENV Variables:
- REGISTRY_URL
- option 1:
https://ui.shadcn.com/r- The URL of the shadcn registry - option 2:
http://localhost:3000>/r- The URL of your own registry server (follow this guide: https://ui.shadcn.com/docs/blocks)
- option 1:
- STYLE - is the style of the shadcn registry -
new-york
npx
You can also run the MCP server using npx, which is convenient for local development or CI/CD pipelines.
Add the following to your mcp.json file:
"shadcn-registry-manager": {
"command": "npx",
"args": [
"@reuvenorg/shadcn-registry-manager"
],
"env": {
"REGISTRY_URL": "https://ui.shadcn.com/r",
"WORKSPACE_DIR": "/Library/Projects/mcp-docker/www",
"STYLE": "new-york"
}
}
Claude
You can also use the MCP server with Claude. Here are the configuration options:
Claude CLI Command:
claude mcp add --scope project shadcn-registry-manager npx @reuvenorg/shadcn-registry-manager -e REGISTRY_URL=https://ui.shadcn.com/r -e WORKSPACE_DIR=/Library/Projects/resume -e STYLE=new-york
Claude .mcp.json Configuration:
{
"shadcn-registry-manager": {
"type": "stdio",
"command": "npx",
"args": [
"@reuvenorg/shadcn-registry-manager"
],
"env": {
"REGISTRY_URL": "https://ui.shadcn.com/r",
"WORKSPACE_DIR": "/Library/Projects/resume",
"STYLE": "new-york"
}
}
}
npx ENV Variables:
- WORKSPACE_DIR - The path to your project directory (mounted as the workspace)
- REGISTRY_URL
- option 1:
https://ui.shadcn.com/r- The URL of the shadcn registry - option 2:
http://localhost:3000/r- The URL of your own registry server (follow this guide: https://ui.shadcn.com/docs/blocks)
- option 1:
- STYLE - is the style of the shadcn registry -
new-york
Example MCP Tools
get_init_instructions: Get project initialization instructionsexecute_init: Run full project initializationget_items: List available registry itemsget_item: Fetch a specific registry itemadd_item: Add a registry item to your projectexecute_add: Add multiple components to your projectget_blocks: Get current blocks from the registry
Example Usage
use the tool `get_blocks` to get the blocks from the registry
use the tool `add_item` - dashboard-01 to your project
Source
made with ❤️ by Reuven Naor
MIT License. See LICENSE for details.
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。