# MCP

Use the shadcn MCP server to discover and install mcpcn components.

> For the complete documentation index, see [llms.txt](/llms.txt). Markdown variants are available by appending `.md` to any URL or sending an `Accept: text/markdown` header. An agent skill is available at [/.well-known/agent-skills/site-skill.md](/.well-known/agent-skills/site-skill.md).



[Model Context Protocol](https://modelcontextprotocol.io/) standardizes how AI assistants connect to tools and data sources. The shadcn CLI includes an MCP server that can browse, search, and install items from compatible registries such as **mcpcn**.

## Configure the registry [#configure-the-registry]

Add the mcpcn namespace to your project's `components.json` file:

```json title="components.json"
{
  "registries": {
    "@mcpcn": "https://mcpcn.dev/r/{name}.json"
  }
}
```

See the [Registry](/docs/registry) page for CLI usage and full-URL installation.

## Configure MCP [#configure-mcp]

Run the shadcn MCP setup command from your project:

```bash
npx shadcn@latest mcp init
```

Choose your MCP client when prompted, finish its setup, and restart the client so it can load the server.

## Usage [#usage]

With MCP enabled, your assistant can discover mcpcn blocks and install them through natural-language requests. Try prompts such as:

* Show me all available components in the mcpcn registry.
* Add `@mcpcn/message-bubble` and `@mcpcn/quick-reply` to my project.
* Build a payment flow using the mcpcn order and payment confirmation blocks.
* Create an event experience with an event card, event details, and ticket selection.
* Find a compact mcpcn component for displaying status and progress.

The MCP server reads registries from `components.json`, then uses the shadcn CLI to copy selected source files and their dependencies into your project.

For client-specific configuration and troubleshooting, see the [shadcn MCP documentation](https://ui.shadcn.com/docs/mcp).