# AI & agents

> fancyborders is built to be found and used by AI coding agents as well as people.

## MCP server

A hosted MCP server, nothing to install: https://fancyborders.vercel.app/mcp (Streamable HTTP, no auth).

```bash
claude mcp add --transport http fancyborders https://fancyborders.vercel.app/mcp
```

Cursor (`.cursor/mcp.json`) and most other clients:

```json
{
  "mcpServers": {
    "fancyborders": {
      "url": "https://fancyborders.vercel.app/mcp"
    }
  }
}
```

Tools: `list_components`, `search_components`, `get_component`, `get_install_command`, `get_component_source`, `get_guide`.

## Agent skill

```bash
npx skills add ommgh/fancyborders
```

The skill is in the repository at `skills/fancyborders`: https://github.com/ommgh/fancyborders/tree/master/skills/fancyborders

## shadcn registry and MCP

```bash
npx shadcn@latest registry add @fancyborders=https://fancyborders.vercel.app/r/{name}.json
npx shadcn@latest mcp init
```

The shadcn MCP server can then browse, search and install from the registry.

## llms.txt and Markdown

- https://fancyborders.vercel.app/llms.txt: an index of the docs.
- https://fancyborders.vercel.app/llms-full.txt: every page in one file.
- Any docs page as Markdown: add `.md` to its address, or request it with `Accept: text/markdown`.
