Mogho AI - CLI tool to orchestrate MCP servers using YAML files

A TypeScript-based CLI tool to parse YAML files and create Model Context Protocol (MCP) servers.

Getting Started with MoghoAI

Using MoghoAI is straightforward and efficient.

1. Initialize a New Project

Begin by initializing a new project with the following command:

npx moghoai init

This command scaffolds a new project with all the necessary packages and file structures required to get started.

2. Configure Your Tools

Next, define your desired configuration rules in the server-config.yaml file. This is where you specify the tools your MCP server will use. Callback functions associated with these tools should be placed in the action directory.

3. Start the Local MCP Server

To start your MCP server locally, run:

npm run start -- [<path-to-yaml-config>]

And that’s it—you now have a local MCP server up and running.

You can then add your MCP server to Cursor for testing. As demonstrated, the two tools (get-alerts and get-forecast) are accessible through the local server. For example, asking Cursor for weather alerts in Austin, Texas triggers the get-alerts tool, and the responses are nicely formatted and returned via the MCP server.

4. Deploy the MCP Server Remotely

To deploy your MCP server to a remote environment, simply run:

npm run deploy

Once deployed, you’ll receive a URL to your remote MCP server. Copy this URL and update your remote MCP integration in Cursor. Cursor will now be connected to your remote server.

5. Test the Remote Integration

Let’s test it again—this time by asking Cursor for the weather forecast for a specific location with:

Latitude: 834943034 Longitude: 3489348823799 Cursor will now use the get-forecast tool via the remote MCP server to return the forecast information.