MCP server basics

SendPulse's MCP server is a tool that helps you manage your account through a chat with an AI assistant. It uses SendPulse API methods to get information and take actions, such as request statistics, run message campaigns, or update user data.

You can connect this MCP server to OpenAI, Claude, Cursor IDE, and more. Choose an environment, and the server will connect it with your SendPulse account.

Let's talk about how the MCP server works and how to interact with it through an AI client.

Overview

The MCP server acts as middleware between your AI assistant and your SendPulse account. It processes requests through the SendPulse API and sends results back to chat, so you can manage everything without leaving the conversation.

Key terms to learn before setup:

• Client is the app where you chat with your AI assistant, such as ChatGPT or Cursor IDE.
• MCP is the middleware between the client and your SendPulse account. It shows available actions and executes them through the API.
• Tools are specific actions you can take in your account.

Once connected, the MCP server operates as follows:

1. You ask your AI assistant something in chat.
2. It forwards your request to the MCP server.
3. The MCP server calls the API to get data or take an action.
4. The AI assistant sends the result back to your chat.

All requests are billed according to your client's API pricing policy.

Add SendPulse’s MCP server

You can add SendPulse's MCP server to various clients. Setup instructions vary by client. To connect the MCP server, you need its link.

To get it, go to Account settings > API and copy the link in the MCP section.

Most clients use basic authentication to recognize and access the supported API methods. Once you add the configuration file, the client redirects you to the browser, where you enter your login credentials.

Here is an example of a configuration file for basic authentication:

{
  "mcpServers": {
    "sp": {
      "url": "https://mcp.sendpulse.com/mcp"
    }
  }
}

Some clients use API credentials instead. In this case, you need a Client ID and Secret. To get them, go to Account settings > API > Client credentials.

Here is an example of a configuration file for OAuth 2.0 authentication:

{
  "servers": {
    "sendpulse": {
      "type": "http",
      "url": "https://mcp.sendpulse.com/mcp",
      "headers": {
        "X-Client-ID": "YOUR_CLIENT_ID",
        "X-Client-Secret": "YOUR_CLIENT_SECRET"
      }
    }
  }
}

Chat with your AI assistant

Once you've connected the MCP server, you can send messages and take other actions in your SendPulse account using tools.

To see all available tools and manage access permissions in your client, go to the Tools section.

To chat with your AI assistant, go to Chats and type in your request using natural language. The AI agent will then show the tool it plans to run. If you have multiple chatbots or subscribers with a similar name, it will ask which one you are referring to.

Here's how request results are returned to your chat:

• If successful, you will receive account information or action results.
• If an error occurs, check your client logs and authentication status.

Last Updated: 29.09.2025