# Lumar MCP Server

https://api-docs.lumar.io/docs/lumar-mcp/overview

# Lumar MCP Server

The Lumar MCP (Model Context Protocol) server connects compatible AI assistants to your live Lumar data. You can investigate crawl results, analyze AI Visibility, and carry out supported actions in natural language without copying data between applications.

Use this remote MCP endpoint:

```
https://mcp.lumar.io/mcp
```

The server uses the standard Streamable HTTP transport. You can add it directly to Claude, ChatGPT, Cursor, or another remote-MCP client, or install the optional [Lumar analytics plugin](./plugins.md) for Claude Code, Cursor, and OpenAI Codex.

## What you can access

The server currently covers two Lumar product surfaces:

- **Lumar Analyze** -- inspect projects and crawls, filter report rows, aggregate URL data, track health trends, manage segments and tasks, run exports and Single Page Requester jobs, create custom metrics, and work with connected Jira projects.
- **Lumar AI Visibility** -- inspect projects, topics, prompts, brands, citations, mentions, visibility scores, content-evaluation results, providers, and Google Search Console bindings; supported write tools can also manage configuration and trigger new runs.

The tools available in a session depend on:

1. The Lumar products and account role available to the authenticated principal.
2. The [toolsets granted to the connection](./authentication-and-toolsets.md).
3. Any action restrictions imposed by the AI client or its subscription plan.

## Quick setup

You need a Lumar account with the **MCP Server** subscription addon and an AI client that supports remote MCP servers.

1. Add a remote MCP server or custom connector in your AI client.
2. Enter `https://mcp.lumar.io/mcp`.
3. Complete the Lumar login and choose the toolsets to grant.

Clients that accept JSON configuration commonly use:

```json
{
  "mcpServers": {
    "lumar": {
      "type": "http",
      "url": "https://mcp.lumar.io/mcp"
    }
  }
}
```

The exact configuration shape varies by client. For current Claude, ChatGPT, and Cursor UI steps, see [How to Get Started with the Lumar MCP Server](https://help.lumar.io/hc/en-us/articles/49064142165521-How-to-Get-Started-with-the-Lumar-MCP-Server). For Claude Code, Cursor, and Codex, the [Lumar analytics plugin](./plugins.md) configures the connection and adds guided workflows.

After connecting, verify access with:

```
Which Lumar accounts and products can I access?
```

For CI, scheduled agents, or another environment without a browser, use [service-account authentication](./authentication-and-toolsets.md#service-account-keys).

## Which Lumar MCP endpoint do I need?

Lumar provides two distinct MCP servers:

| Endpoint                        | Use it for                                                                        | Authentication |
| ------------------------------- | --------------------------------------------------------------------------------- | -------------- |
| `https://mcp.lumar.io/mcp`      | Query live Lumar product data and perform supported Analyze/AI Visibility actions | Required       |
| `https://api-docs.lumar.io/mcp` | Search and read this API documentation                                            | Not required   |

This section documents the first endpoint. If you only want an assistant to search the API reference, see [MCP Server for API Docs](../mcp-server.md).

## Recommended path

1. Connect directly using the quick setup above, or install the [Lumar analytics plugin](./plugins.md).
2. Review [Authentication and toolsets](./authentication-and-toolsets.md), especially before enabling write actions or headless automation.
3. Read [Usage and troubleshooting](./usage.md) for reliable scoping, pagination, asynchronous work, and common connection issues.

:::caution
Your Lumar data passes through the AI provider you connect. Review that provider's privacy and data-training terms, and follow your organization's security policy before connecting production accounts.
:::

## Related resources

- [Lumar MCP Server release note](https://help.lumar.io/hc/en-us/articles/49064047611409-June-2026-Lumar-MCP-Server)
- [How to Get Started with the Lumar MCP Server](https://help.lumar.io/hc/en-us/articles/49064142165521-How-to-Get-Started-with-the-Lumar-MCP-Server)
- [Lumar MCP getting-started article](https://www.lumar.io/blog/company-news/how-to-get-started-with-the-lumar-mcp-server/)
- [Lumar plugins on GitHub](https://github.com/deepcrawl/lumar-plugins)