Openai MCP Implementation

A simple server project that interacts with AI models of OpenAI and Claude as a master control program (MCP), providing a unified API interface and automatic provider detection function.

Artificial intelligence chatbots Developer tools #AI Gateway #Multi - model #API Aggregation #Master Control Service .JavaScript

rating : 2 points

downloads : 22

update time : 2025-04-28

What is the MCP AI Gateway?

MCP (Master Control Program) is an intelligent proxy server that can automatically identify and call the most suitable AI model of OpenAI or Anthropic (Claude) to process your requests. It simplifies the calling process of APIs from different AI providers and provides a unified interface.

How to use the MCP service?

Simply install the service and send a request containing your question, and MCP will automatically select the best AI model for processing. You can also specify a particular provider or model.

Applicable scenarios

Suitable for developers who need to use multiple AI models simultaneously, teams that want to simplify API calls, and application scenarios that require intelligent routing of AI requests.

Main Features

Intelligent RoutingAutomatically select the API of OpenAI or Claude based on the model name without manual switching

Unified API InterfaceProvide standardized request/response formats for different AI providers

Multi - model SupportSupport the latest models of OpenAI and Claude, including the GPT - 4 and Claude - 3 series

Built - in Test InterfaceProvide a web interface to directly test API functions

Advantages and Limitations

Advantages

Simplify the integration process of multiple AI providers

Automatically select the optimal model to process requests

Unify error handling and log recording

Built - in test tools for quick function verification

Limitations

Need to deploy and maintain the server by yourself

Depend on third - party API keys and quotas

New models require updating the service version

How to Use

Install the Service

Clone the code repository and install dependencies

Configure API Keys

Copy the example environment file and add your OpenAI and Claude API keys

Start the Service

Run the service (default port 3000)

Send a Request

Send a request through the API endpoint or use the built - in test interface

Usage Examples

Automatically select an AI to handle technical problemsLet MCP automatically select the most suitable AI model to answer complex technical questions

Quick content summarySpecify to use a fast model to handle simple tasks

Multi - model comparison testSend the same question to different models to compare the results

Frequently Asked Questions

How can I know which AI model processed my request?

Can I use multiple AI models simultaneously?

Does the service support streaming responses?

How to add a new AI provider?

Related Resources

OpenAI API Documentation

Official OpenAI API reference documentation

Claude API Documentation

Official Anthropic Claude API documentation

GitHub Repository

Project source code

Node.js Installation Guide

Node.js environment required to run the service

🚀 MCP Implementation: Master Control Program for OpenAI and Claude

A simple server acting as a Master Control Program (MCP) to interact with AI models from OpenAI and Anthropic (Claude).

🚀 Quick Start

This project provides a straightforward way to interact with OpenAI and Claude models through a unified interface. Follow the steps below to get started.

📦 Installation

# Clone the repository
git clone https://github.com/Spysailor/openai-mcp-implementation.git

# Navigate to the directory
cd openai-mcp-implementation

# Install dependencies
npm install

# Configure environment variables
cp .env.example .env
# Next, edit the .env file and add your API keys

✨ Configuration

Edit the .env file to add your API keys:

# MCP configuration
PORT=3000
OPENAI_API_KEY=Fill in your OpenAI API key here
ANTHROPIC_API_KEY=Fill in your Claude API key here

💻 Usage Examples

Basic Usage

# Start the server
npm start

The server will run at http://localhost:3000.

Advanced Usage

The server provides multiple API endpoints for different use cases.

POST /api/mcp

This is a unified endpoint that automatically detects the provider based on the requested model.

Request Parameters (JSON)

prompt (required): The text content to send to the AI.
model (optional): The model to use (default: 'gpt-4' for OpenAI).
maxTokens (optional): The maximum number of tokens for the response (default: 2000).
provider (optional): Specify the provider ('openai', 'claude', or 'auto' for automatic detection).

Request Example

curl -X POST http://localhost:3000/api/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Please explain in Chinese how artificial intelligence works, divided into three paragraphs.",
    "model": "gpt-4",
    "maxTokens": 500,
    "provider": "auto"
  }'