Skip to content

Getting Started

This guide walks you through installing and configuring the Apitomy Data Models MCP server with your AI coding assistant.

Prerequisites

  • Node.js 18 or later
  • An MCP-compatible AI coding assistant (e.g., Claude Code, Cursor, Windsurf)

Installation

Option 1: Run with npx (no install)

The simplest way — no global installation required:

npx @apitomy/data-models-mcp

Option 2: Global install

npm install -g @apitomy/data-models-mcp

Then run with:

apitomy-data-models-mcp

Configuration

Claude Code

Add the MCP server to your Claude Code settings. Create or edit .claude/settings.json:

{
  "mcpServers": {
    "apitomy-data-models": {
      "command": "npx",
      "args": ["-y", "@apitomy/data-models-mcp"]
    }
  }
}

Cursor

Add to your Cursor MCP configuration (.cursor/mcp.json):

{
  "mcpServers": {
    "apitomy-data-models": {
      "command": "npx",
      "args": ["-y", "@apitomy/data-models-mcp"]
    }
  }
}

VS Code (Copilot)

Add to your VS Code settings (.vscode/settings.json):

{
  "mcp.servers": {
    "apitomy-data-models": {
      "command": "npx",
      "args": ["-y", "@apitomy/data-models-mcp"]
    }
  }
}

Verifying the Setup

Once configured, ask your AI assistant to test the connection:

"Create a new OpenAPI 3.1 document called 'test' with title 'My API'"

The assistant should use the document_create tool to create the document and confirm it was created successfully.

Basic Workflow

Here's a typical workflow for creating an API specification:

1. Create a document

"Create a new OpenAPI 3.1 document called 'petstore' with title 'Petstore API' version '1.0.0'"

2. Add paths and operations

"Add a GET /pets operation that returns a list of pets"

3. Add schemas

"Add a Pet schema with id (integer), name (string), and status (string) properties"

4. Validate

"Validate the petstore document"

5. Export

"Export the petstore document as YAML"

Loading an Existing Document

You can also load an existing OpenAPI or AsyncAPI document:

"Load the file openapi.json into a session called 'myapi'"

The MCP server will parse the document and make it available for querying and editing.

Next Steps