> For the complete documentation index, see [llms.txt](https://docs.exads.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.exads.com/mcp/mcp-connecting-ai-client.md).

# Connecting AI Client

## Prerequisites

Before connecting, you need an active account on the platform. The MCP add-on must also be enabled for the network.

## Authentication <a href="#authentication" id="authentication"></a>

The MCP Server uses OAuth to authenticate your connection. When you add the connector to your AI client, you will be redirected to the admin panel to authorize access. Once authorized, the connection is established, and you can start working immediately - no need to manually manage API keys or tokens within your AI tool.

**Your credentials are never shared with or stored by the AI assistant.** The MCP Server handles authentication securely on your behalf.

## Choosing what the connection can do

After you authenticate, and before you are returned to your AI client, you are asked to choose what this connection is allowed to do. Each capability your account holds is shown with an access level of Read, Write, or No access. Everything defaults to Read.

These settings apply to that connection only, and never grant more than your account already permits. See [Available tools](/mcp/mcp-tools.md) for what each capability covers.

To change the access levels later, reconnect the AI client and choose again. Reconnecting replaces the previous connection immediately, so the old access stops working straight away.

Connections authorized before this feature was introduced are read-only until you reconnect.

## Connecting your AI client <a href="#connecting-your-ai-client" id="connecting-your-ai-client"></a>

The MCP Server URL is:

```
https://mcp.<ad-network-domain>/mcp
```

Below are setup instructions for the most popular AI clients.

### Claude.ai and Claude Desktop <a href="#claudeai-and-claude-desktop" id="claudeai-and-claude-desktop"></a>

Claude supports MCP connectors natively. To connect:

1. Open Claude, go to the **Customize** section and click **Connectors**.
2. Click the **+** icon to add a connector, then choose **Add Custom Connector**.
3. Enter a name and the URL: `https://mcp.<ad-network-domain>/mcp`. The connector will be added to the list.
4. Click **Connect**. You will be redirected to the admin panel to authorize access.
5. Once authorized, you will be returned to Claude and the connection will be active.

### Claude Code <a href="#claude-code" id="claude-code"></a>

If you use Claude Code from the command line, you can add the MCP server with a single command:

```bash
claude mcp add <ad-network> https://mcp.<ad-network-domain>/mcp --transport http
```

The first time you use the server in a session, Claude will automatically prompt you to authorize via the admin panel. You can also authorize manually before first use:

1. Start an interactive session by running `claude`.
2. Run the `/mcp` command.
3. Select the server.
4. Click **Authorize**.

### ChatGPT <a href="#chatgpt" id="chatgpt"></a>

ChatGPT supports MCP connectors through its Plugins. To connect:

1. Open ChatGPT, go to **Settings**, then **Plugins**, and then **MCPs**.
2. Click **+Add Server**, then enter a name.
3. Choose **Streamable HTTP** and add the URL: `https://mcp.<ad-network-domain>/mcp`
4. Click **Save**, go back to the Plugin menu, then click **Authenticate**. You will be redirected to the admin panel to authorize access.
5. Once authorized, you will be returned to ChatGPT and the connection will be active.

***

{% hint style="info" %}
There are a couple of differences if you are using the Browser app:&#x20;

* You need to enable **Developer Mode** in order to add a custom MCP. You can enable this setting through the **Security & Login** settings.
* MCPs are added directly through the Plugin menu by pressing the **+** icon.
  {% endhint %}

***

### Cursor <a href="#cursor" id="cursor"></a>

Cursor supports MCP connectors through its customize menu. To connect:

1. Open Cursor, go to **Customize**, click **MCPs** and then choose **New MCP Server**.
2. Add the following object to your list of MCPs in the **mcp.json** file:

```json
"<ad-network>": {
      "url": "https://mcp.<ad-network-domain>/mcp"
    }
```

3. Save the file. The list of MCPs will show the newly added connector. Click the **Authorize** button to be redirected to the admin panel and authorise access.
4. Once authorised, you will be returned to Cursor and the connection will be active.

### Other AI clients <a href="#other-ai-clients" id="other-ai-clients"></a>

| Field          | Value                                 |
| -------------- | ------------------------------------- |
| Server URL     | `https://mcp.<ad-network-domain>/mcp` |
| Transport      | Streamable HTTP                       |
| Authentication | OAuth (redirects to the admin panel)  |

Refer to your AI client's documentation for instructions on adding a custom MCP connector.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.exads.com/mcp/mcp-connecting-ai-client.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.