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

# MCP Overview

> Learn about the Model Context Protocol (MCP) and how to extend emma's capabilities with external tools.

## What is MCP?

The **Model Context Protocol (MCP)** is an open standard that enables AI applications to securely connect to external data sources and tools. Think of MCP like a USB-C port for AI applications — a standardized way to connect AI apps to different services.

With MCP, you can give AI agents like Cursor or Claude Code access to emma's tools, integrations, and saved Skills through a unified interface — directly from your IDE or terminal.

## Why Use MCP with emma?

By connecting your AI coding agent to the emma MCP Server, you can:

<CardGroup cols={2}>
  <Card title="Access 100+ Integrations" icon="puzzle-piece">
    Search and execute any of emma's Slack, GitHub, Notion, Gmail, and other integrations without leaving your editor
  </Card>

  <Card title="Reuse Saved Skills" icon="book">
    Save a workflow once in emma and reuse it from any machine, IDE, or agent
  </Card>

  <Card title="Automate OAuth" icon="key">
    emma handles OAuth connections for you — tools return a connect link when authentication is needed
  </Card>

  <Card title="Standardized Protocol" icon="plug">
    Use the same protocol across different AI applications
  </Card>
</CardGroup>

## Supported Transport

emma MCP Server is a **remote, hosted MCP server** that speaks **Streamable HTTP**. There's nothing to install — your client connects directly over HTTPS using your emma API key.

| Transport           | Status                      | Description                                                                                                                                                                   |
| ------------------- | --------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Streamable HTTP** | Recommended                 | Remote server hosted by emma at `/api/v1/mcp`. Just add a URL and an API key — no local process to install or keep running                                                    |
| **stdio**           | Supported (local, optional) | For clients that can't speak remote HTTP MCP yet, emma also publishes a local stdio bridge (`@duzzle/emma-mcp` on npm) that proxies the same tools over your existing API key |
| SSE                 | Not supported               | Server-Sent Events                                                                                                                                                            |

<Note>
  The remote Streamable HTTP server exposes exactly the same tools as the local stdio bridge, so you can switch between them without changing how you use emma from your agent.
</Note>

## Supported Clients

Any MCP client that supports the Streamable HTTP transport (with custom headers for authentication) can connect to emma. emma officially documents setup for:

* **Cursor** — AI-powered code editor
* **Claude Code** — Anthropic's coding agent CLI
* **Codex** — OpenAI's coding agent CLI

<Note>
  Clients that only support the stdio transport (or don't yet support custom HTTP headers) can use the local `@duzzle/emma-mcp` package instead — see the [Setup Guide](/mcp/setup) for details.
</Note>

## Next Steps

<CardGroup cols={2}>
  <Card title="emma MCP Server" icon="server" href="/mcp/emma-mcp-server">
    Learn about emma's official MCP server and available tools
  </Card>

  <Card title="Setup Guide" icon="wrench" href="/mcp/setup">
    Get started with configuring emma MCP in your environment
  </Card>
</CardGroup>