Skip to Content
The Makinari API repo now includes an MCP Server — connect your AI models directly. View on GitHub →
Get startedPath: MCP in 5 Minutes

MCP in 5 Minutes

The Model Context Protocol (MCP) path is for developers who want to expose Makinari’s Agent tools to their local AI assistants (like Cursor, Claude Desktop, or a custom agent).

Goal: Connect your Cursor IDE to the Makinari MCP Server to allow your local AI to interact with your Makinari Site.

Prerequisites

  • A Makinari account and Site ID.
  • An API Key (see First API Request).
  • Cursor IDE installed.

Step 1: Configure Cursor

Cursor natively supports adding remote MCP servers.

  1. Open Cursor Settings (gear icon).
  2. Navigate to Features > MCP Servers.
  3. Click + Add New MCP Server.
  4. Configure the server:
    • Type: sse
    • Name: Makinari
    • URL: https://backend.makinari.com/api/mcp

Step 2: Add Authentication Headers

The Makinari MCP server requires your API key and Site ID. Cursor allows you to pass these as custom headers.

Add the following to the Headers JSON configuration in Cursor:

{ "Authorization": "Bearer YOUR_API_KEY", "x-mcp-site-id": "YOUR_SITE_ID" }

Replace YOUR_API_KEY and YOUR_SITE_ID with your actual values.

Step 3: Test the integration

  1. Save the MCP server configuration. Cursor should display a green indicator showing it successfully connected and loaded the tools.
  2. Open the Cursor AI Chat (Cmd/Ctrl + L).
  3. Type a prompt that requires Makinari context, for example:

    “Use the Makinari tools to list the latest 5 leads from my site.”

  4. Cursor will automatically call the list_leads tool via the MCP server and summarize the results for you.

Next steps

Last updated on