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

# MCP Server Overview

> SmartAlex MCP server lets AI assistants like Claude, ChatGPT, Gemini, Cursor, and VS Code manage voice agents, campaigns, contacts, and analytics through the Model Context Protocol.

Connect SmartAlex to ChatGPT, Claude, Cursor, and other AI assistants, then run your entire voice platform just by chatting: *"create an agent for my clinic,"* *"start the follow-up campaign,"* *"show me yesterday's calls."*

<CardGroup cols={3}>
  <Card title="ChatGPT" icon="https://mintcdn.com/smartalex/sirBslEh6uou56ET/images/clients/openai.svg?fit=max&auto=format&n=sirBslEh6uou56ET&q=85&s=b8b713760189010df248c4799c829b43" href="/mcp/getting-started#chatgpt" width="24" height="24" data-path="images/clients/openai.svg">
    Hosted cloud endpoint. No install.
  </Card>

  <Card title="Claude" icon="https://mintcdn.com/smartalex/sirBslEh6uou56ET/images/clients/claude.svg?fit=max&auto=format&n=sirBslEh6uou56ET&q=85&s=a99573f3e7e3b68a1ce410af0db5b0c6" href="/mcp/getting-started#claude-desktop" width="24" height="24" data-path="images/clients/claude.svg">
    Claude Desktop and Claude web, in two minutes.
  </Card>

  <Card title="Claude Code" icon="https://mintcdn.com/smartalex/sirBslEh6uou56ET/images/clients/anthropic.svg?fit=max&auto=format&n=sirBslEh6uou56ET&q=85&s=7ecbd8c37c36f9c89a95108444d90b9d" href="/mcp/getting-started#claude-code" width="24" height="24" data-path="images/clients/anthropic.svg">
    Cloud or local, one command.
  </Card>

  <Card title="Cursor" icon="https://mintcdn.com/smartalex/sirBslEh6uou56ET/images/clients/cursor.svg?fit=max&auto=format&n=sirBslEh6uou56ET&q=85&s=983cf68e9423b5028bee275a30785f18" href="/mcp/getting-started#cursor" width="24" height="24" data-path="images/clients/cursor.svg">
    Add to your `.cursor/mcp.json`.
  </Card>

  <Card title="Gemini" icon="https://mintcdn.com/smartalex/sirBslEh6uou56ET/images/clients/googlegemini.svg?fit=max&auto=format&n=sirBslEh6uou56ET&q=85&s=b5fd8bc0763ed6fc0ad30d19ff75e657" href="/mcp/getting-started#gemini-cli" width="24" height="24" data-path="images/clients/googlegemini.svg">
    Gemini CLI via `settings.json`.
  </Card>

  <Card title="Any MCP client" href="/mcp/getting-started">
    VS Code, Windsurf, and any MCP-compatible app.
  </Card>
</CardGroup>

<Card title="New here? Connect in under 2 minutes" href="/mcp/getting-started" horizontal>
  The Getting Started guide walks you through it step by step. No coding needed for the cloud option.
</Card>

## What is MCP?

The Model Context Protocol is an open standard that connects AI assistants to external tools and data. Think of it as a universal adapter between your AI and SmartAlex. Instead of switching between dashboards, you just ask your AI assistant to manage contacts, launch campaigns, analyze calls, and track deals, all through conversation.

MCP is supported by Claude, Claude Code, Cursor, VS Code (GitHub Copilot), ChatGPT, Gemini CLI, and a growing ecosystem of AI clients.

## What you can do

The SmartAlex MCP server exposes **28 tools**, **7 resources**, and **5 prompts**, giving AI assistants full operational access to your SmartAlex workspace. Your AI client sits on one side, your workspace on the other, and MCP connects them across eight domains.

```mermaid theme={null}
flowchart LR
    subgraph Clients["Your AI client"]
      C1["Claude"]
      C2["ChatGPT"]
      C3["Cursor"]
      C4["Gemini"]
    end
    Clients <-->|"MCP"| M(["SmartAlex MCP server"])
    M --> D1["Contacts"]
    M --> D2["Agents"]
    M --> D3["Phone numbers"]
    M --> D4["Campaigns"]
    M --> D5["Calls"]
    M --> D6["Deals"]
    M --> D7["Webhooks"]
    M --> D8["Platform and Widgets"]
```

At a glance, you can:

* **Manage contacts**: list, create, update, and organize your contact database
* **Create voice agents**: build an agent from a website URL or fully customize it, with automatic voice sync
* **Run campaigns**: create campaigns, add contacts, start calling, and track progress
* **Analyze calls**: browse call history and read full transcripts with sentiment data
* **Track deals**: manage your sales pipeline from lead to closed
* **Set up webhooks**: create event-driven integrations with external systems
* **Monitor the platform**: check health, usage stats, and system status

## Tools (28)

SmartAlex exposes 28 tools across 8 domains, all prefixed `smartalex_` so they never clash with other MCP servers.

| Domain             | Tools | What it covers                                     |
| ------------------ | ----- | -------------------------------------------------- |
| Contacts           | 4     | List, get, create, and update contacts             |
| Agents             | 5     | List, get, create, update, and delete voice agents |
| Phone numbers      | 3     | List, search, and buy numbers                      |
| Campaigns          | 5     | Create, populate, and start outbound campaigns     |
| Calls              | 2     | List calls and read full transcripts               |
| Deals              | 4     | Manage the sales pipeline                          |
| Webhooks           | 3     | Subscribe to and manage event webhooks             |
| Platform & Widgets | 2     | Platform status and widget install code            |

<Card title="Full tools reference" href="/mcp/tools-reference" horizontal>
  Every tool with its parameters, examples, and responses.
</Card>

## Resources and prompts

Beyond tools, the server ships **7 read-only resources** that give your AI client ambient context about your workspace, and **5 prompts** that are ready-made workflows (campaign strategy, lead qualification, call analysis, and more).

<Card title="Resources & prompts" href="/mcp/resources-and-prompts" horizontal>
  The 7 context resources and 5 workflow prompts, and how to use them together.
</Card>

## Supported AI clients

The SmartAlex MCP server works with any MCP-compatible client:

| Client                       | Transport                     | Status                          |
| ---------------------------- | ----------------------------- | ------------------------------- |
| **Claude Code**              | HTTP (cloud) or stdio (local) | Fully supported                 |
| **Claude Desktop**           | stdio (local)                 | Fully supported                 |
| **Cursor**                   | stdio (local)                 | Fully supported                 |
| **VS Code (GitHub Copilot)** | stdio (local)                 | Fully supported                 |
| **Windsurf**                 | stdio (local)                 | Fully supported                 |
| **Gemini CLI**               | stdio (local)                 | Fully supported                 |
| **ChatGPT**                  | HTTP (cloud)                  | Supported via cloud endpoint    |
| **Custom clients**           | HTTP or stdio                 | Any MCP-compatible client works |

## Connect and authenticate

The server runs two ways: a hosted **cloud** endpoint you connect to by URL, or a **local** install over stdio. Both authenticate with an API key or OAuth.

<CardGroup cols={3}>
  <Card title="Getting started" href="/mcp/getting-started">
    Connect a client in under two minutes.
  </Card>

  <Card title="Authentication" href="/mcp/authentication">
    API keys, OAuth, scopes, and rate limits.
  </Card>

  <Card title="Architecture" href="/mcp/architecture">
    Cloud vs local, and how it all fits.
  </Card>
</CardGroup>

## Next steps

* [Getting Started](/mcp/getting-started), connect SmartAlex to your AI assistant in under 2 minutes
* [Local install (npm)](/mcp/npm-package), run it locally with Claude Desktop, Cursor, VS Code, or Windsurf
* [Tools reference](/mcp/tools-reference), all 28 tools with parameters and examples
* [Resources & prompts](/mcp/resources-and-prompts), the 7 resources and 5 workflow prompts
* [Usage examples](/mcp/examples), real-world workflows and conversation examples
* [Error reference](/mcp/errors), error codes and troubleshooting