Skip to Content
The Makinari API repo now includes an MCP Server — connect your AI models directly. View on GitHub →
MCP ServerSetupInstall Locally

Install Locally

If you prefer to run the MCP server on your own machine, follow these steps.

Prerequisites

First, clone the repository:

git clone https://github.com/Makinari/API.git cd API

Ensure you have the project dependencies installed.

npm install

Configuration

The MCP server requires the following environment variables:

  • MCP_SITE_ID (required): The ID of your site.
  • MCP_USER_ID (optional): The ID of the user.
  • MCP_INSTANCE_ID (optional): The ID of the instance (default: ‘default’).
  • API_URL (optional): The URL of the API server. Defaults to http://localhost:3000.
    • You can use the Makinari API (if you have access) or run the API locally from this repository.
    • If running locally, ensure the API server is started (e.g., npm run dev) and reachable at this URL.
  • REST_API_KEY (optional): Required if using a remote API (e.g., Makinari API).
  • SERVICE_API_KEY (optional): Required if running the API locally. This should be set in the .env file of the API project.

Running the Server

Stdio Mode

To run the MCP server in stdio mode (for local development or Cursor integration):

npm run mcp

HTTP Mode

The MCP server is also exposed via an HTTP endpoint for remote clients.

POST /api/mcp

This endpoint handles JSON-RPC requests over HTTP.

Last updated on