Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.karada.ai/llms.txt

Use this file to discover all available pages before exploring further.

Welcome to Karada! This tutorial walks you through the core flow of the platform: creating a project and transforming a standard REST API into an AI-ready Model Context Protocol (MCP) server. What you’ll learn:
  • How to create a project in your workspace
  • How to configure an Auto-MCP integration
  • How to generate and deploy your MCP server

​
Prerequisites

Before you begin, ensure you have:
  • Created an account on Karada.ai.
  • A valid OpenAPI or Swagger URL (e.g., https://api.example.com/openapi.json). If you don’t have one, use a public API like the Petstore API.

​
Step 1: Create a Workspace and Project

A project serves as the top-level container for all your deployments, integrations, and usage tracking.
  1. Navigate to your Dashboard.
  2. If you haven’t already, create a new Team Workspace.
  3. Click the Create Project button in the top right corner.
  4. Select the Auto-MCP template from the list of available providers.
  5. Provide your project with a descriptive name (e.g., Petstore AI Agent).

​
Step 2: Configure Auto-MCP

Karada’s Auto-MCP feature converts any API into an intelligent tool.
1

Select Specification Type

Choose between OpenAPI or Swagger depending on the format of your API specification. Karada supports both widely-used standards.
2

Provide the API URL or File

Paste the URL pointing to your .json or .yaml specification file. Alternatively, upload the file directly from your local machine.
3

Review the Detected Tools

Karada parses the specification immediately. Review the list of detected endpoints, required parameters, and descriptions. These act as the exact tools your AI assistant will use.

​
Step 3: Generate and Deploy

Once you review and approve the detected tools:
  1. Click the Generate Server button.
  2. Wait a few seconds while Karada automatically generates the MCP server.
  3. Choose your preferred hosting method:
    • Download as ZIP: Run the server locally on your machine.
    • Deploy to Karada: Host the server instantly on the Karada infrastructure.

​
Running Locally (Downloaded ZIP)

If you download the server to run locally, follow these steps to start it: 
# Unzip the downloaded folder
unzip karada-mcp-server.zip
cd karada-mcp-server

# Install dependencies
npm install

# Start the generated MCP server
npm start

​
Next Steps

Now that you have your MCP server running, connect it to AI clients like the Claude Desktop App or Cursor IDE. 👉 Learn how to configure AI Clients in our Auto-MCP Deep Dive