> ## Documentation Index
> Fetch the complete documentation index at: https://www.citedy.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP Integration

> Connect to Citedy via Model Context Protocol

Citedy exposes a [Model Context Protocol](https://modelcontextprotocol.io/) (MCP) server that lets MCP-compatible AI clients use Citedy tools natively.

## Connecting

Add to your MCP config (`~/.claude/claude_desktop_config.json` or project `.mcp.json`):

```json theme={null}
{
  "mcpServers": {
    "citedy": {
      "url": "https://mcp.citedy.com/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}
```

## Connection Details

* URL: `https://mcp.citedy.com/mcp`
* Auth: `Authorization: Bearer citedy_agent_...`
* Transport: Streamable HTTP MCP

## Response Envelope

Every tool returns a structured envelope with status and metadata. Async tools include polling instructions (`operation.poll_tool`).

## Available Tools

See the [full MCP tools reference](/mcp/tools) for all **71 tools** with endpoints and descriptions.

Need prebuilt install recipes? See [Skills overview](/skills/overview).