MCP Server Setup Guide

How to connect the Datastreamer MCP Server

MCP or "Model Context Protocol" is a protocol to allow systems and agents to easily interact. Datastreamer provides a number of Agent-ready tools to support product features. This guide will walk you through the connection of your systems to your data pipelines via MCP.

1. Requirements for your MCP Client

To access the Datastreamer MCP server, your MCP client must meet the following requirements:

Register an MCP Server using HTTP transport type.
1. We do not support stdio communication.
Be able to include a long‑lived API key in a custom header called apikey
Your MCP Client should be able to support Tools.

If you do not have a MCP Client, a list of applications that support MCP integrations is available here: https://modelcontextprotocol.io/clients

2. Connecting Datastreamer's MCP Server to your MCP Client

Most MCP Clients allow you to configure servers in a JSON or YAML file.

Configuring using JSON

Below is a generic JSON example (adjust the syntax depending on your MCP client).

{
  "servers": [
    {
      "name": "mcp-server-datastreamer",
      "transport": "http",
      "url": "https://mcp.platform.datastreamer.io",
      "headers": {
        "apikey": "<YOUR_LONG_LIVED_API_KEY>"
      }
    }
  ]
}

📘
How to get your API key
The API Key can be generated within the Datastreamer Portal, on Keys & Secrets page.

Custom Client

If you are writing your own client, you can follow the MCP protocol standards at: https://modelcontextprotocol.io/, you will need to have the following as a minimum:

Your client must use standard HTTP requests,
You must include the apikey header.
The base URL is: https://mcp.platform.datastreamer.io