MCPhy

You can turn your REST API specifications into a conversational MCP server that you interact with in plain language. This enables natural dialogue around your API, while still calling real backend endpoints and returning actual responses. You gain a chat UI, smart guidance when information is missing, and an easy way to export a standalone package for deployment.

How to use

To use the MCP server, follow these practical steps to convert your API spec into a chat-driven experience. You will initialize with your API specification, start the conversational server, and then chat with it to perform real API calls using natural language.

Basic workflow with the MCP client:

• Initialize with your API spec
• Start the conversational server
• Open the web interface and start chatting in plain English to explore endpoints and run actions on your backend.

mcphy init

mcphy serve

open http://localhost:3000

How to install

Prerequisites you need before installing:
- Node.js 18.0.0 or newer
- npm 9.0.0 or newer
- An API specification file in OpenAPI/Swagger or Postman format
- An OpenAI API key if you want AI-powered enhancements (optional)

Install the MCP server globally or try it directly with npx, then verify help outputs to understand usage.

# Option 1: Install globally
npm install -g mcphy

# Option 2: Try without installing
npx mcphy --help

Additional notes

Configuration revolves around providing an API specification in a compatible format. MCPhy supports OpenAPI 3.0+, Swagger 2.0, and Postman Collections (with optional AI-assisted parsing). You can enable AI-powered API understanding for deeper analysis and richer endpoint descriptions when you provide an API key.

Typical usage with an example API file:

• Initialize with your API spec: mcphy init -f path/to/your/api.yaml
• Start the server: mcphy serve
• Open the web interface at http://localhost:3000 and begin chatting with natural language prompts such as “Show me all pets” or “Create a booking for John Doe.”

Available tools

Natural Language Interface

Allows you to ask questions in plain English and translates them into API calls.

Real API Calls

Performs actual requests to your backend, returning real responses.

Smart Guidance

Provides guidance when information is missing or ambiguous to help complete the request.

Web UI

Presents a professional chat interface with real-time responses for a smooth user experience.

Export Package

Exports a standalone package that can be deployed independently.

AI-Powered Understanding (Optional)

Uses an AI model to enhance API understanding, descriptions, and examples when an API key is provided.