Overview

Model Context Protocol revolutionizes how AI agents interact with tools, data sources, and services by providing a standardized interface that ensures security, controllability, and extensibility.

Core Concepts

Model Context Protocol (MCP) is a standardized way for Large Language Models (LLMs) to securely access tools and data sources. It provides a flexible and controlled interface that allows AI models to interact with various services while maintaining security and control.

How MCP Works

Benefits of MCP

Getting Started with Local Installation

Running your own MCP servers locally gives you complete control over your data and functionality. Follow these steps to set up a local MCP environment:

Prerequisites

• Node.js 18+ or Python 3.8+
• Package manager (npm, yarn, pip)
• Basic knowledge of JavaScript/TypeScript or Python

import { MCPServer } from '@model-context-protocol/sdk';

// Create a new MCP server instance
const server = new MCPServer({
  id: 'my-mcp-server',
  name: 'My MCP Server',
  description: 'My first MCP server implementation'
});

// Register a simple tool
server.registerTool({
  id: 'hello',
  name: 'Hello Tool',
  description: 'Say hello to someone',
  parameters: {
    type: 'object',
    properties: {
      name: {
        type: 'string',
        description: 'The name to say hello to',
      },
    },
    required: ['name'],
  },
  handler: async (params) => {
    const { name } = params;
    return { result: `Hello, ${name}!` };
  },
});

// Start the server on port 3000
server.listen(3000, () => {
  console.log('MCP server running on http://localhost:3000');
});

Advanced Local MCP Features

Hosted MCP Platform

Our hosted MCP platform provides a managed environment where you can deploy and run your MCP servers without worrying about infrastructure, scaling, or maintenance. This is the quickest way to get started with MCP.

Getting Started with Hosted MCP

1. Create an Account: Sign up for our hosted MCP platform to get your API key and access to our dashboard.
   Sign Up for Hosted MCP
2. Create a Project: From the dashboard, create a new MCP project with a name and description.
3. Deploy Your MCP Server: Upload your MCP server code or use our templates to create a new server.
4. Configure Authentication: Set up authentication methods and access controls for your MCP server.
5. Start Using Your MCP: Integrate your hosted MCP server with your application using the provided endpoints.

import { MCPClient } from '@model-context-protocol/client';

// Initialize the client with your MCP server URL and API key
const client = new MCPClient({
  serverUrl: 'https://api.mcp.4agent.ai/v1/your-project-id',
  apiKey: 'YOUR_API_KEY',
  timeout: 30000 // optional timeout in milliseconds
});

// Call a tool on the MCP server
async function greetUser() {
  try {
    const response = await client.invokeTool('hello', {
      name: 'World'
    });
    
    console.log(response.result); // Output: "Hello, World!"
    
    return response;
  } catch (error) {
    console.error('Error calling tool:', error);
    throw error;
  }
}

greetUser();

Hosted MCP Features

Understanding the MCP Catalog

Our MCP catalog contains hundreds of ready-to-use MCP servers for various domains and functionalities. From data processing to external API integrations, you can find MCP servers for almost any task.

MCP Categories

How to Choose the Right MCP

Selecting the right MCP servers for your application depends on various factors. Here's a guide to help you make the best choice:

Popular MCP Combinations

Certain MCP combinations work well together for specific use cases. Here are some recommended configurations: