What is MCP?

The Model Context Protocol (MCP) is an open standard for connecting AI applications to external tools and data sources.

The Problem

AI assistants are incredibly capable, but they're limited to the knowledge they were trained on. To be truly useful, they need to:

  • Access your databases and files
  • Interact with your APIs and services
  • Execute actions in your tools (GitHub, Slack, etc.)
  • Query real-time data

Before MCP, every AI application had to build custom integrations for each external service. This led to fragmented, inconsistent, and hard-to-maintain systems.

The Solution: A Standard Protocol

MCP provides a universal way for AI applications to communicate with external services. Instead of custom integrations, you have:

MCP Hosts (Clients)

AI applications like Claude Desktop, IDEs, or custom agents that want to use external tools.

MCP Servers

Services that expose tools, resources, and prompts to MCP hosts. Examples: GitHub, databases, file systems.

Any MCP-compatible host can connect to any MCP server without custom code.

Core Concepts

Tools

Functions that AI can call to perform actions. Each tool has a name, description, and schema defining its inputs. 

{
  "name": "create_issue",
  "description": "Create a GitHub issue",
  "inputSchema": {
    "type": "object",
    "properties": {
      "title": { "type": "string" },
      "body": { "type": "string" },
      "repo": { "type": "string" }
    },
    "required": ["title", "repo"]
  }
}

Resources

Data that can be read by the AI, like files, database records, or API responses. Resources are identified by URIs and can be fetched on demand.

Prompts

Pre-defined prompt templates that servers can expose. These help structure interactions and ensure consistent behavior.

Transport Protocols

MCP supports multiple ways to connect hosts and servers:

stdio (Standard I/O)

For local servers running as subprocesses. Fast and simple, commonly used with Claude Desktop.

SSE (Server-Sent Events)

For remote servers over HTTP. This is what MCP Gateway Pro uses.

WebSocket

For bidirectional real-time communication. Useful for servers that need to push updates.

Where MCP Gateway Fits

MCP Gateway Pro sits between your AI applications and multiple MCP servers, providing:

  • - Single endpoint - Connect to one gateway instead of managing multiple server connections
  • - Server aggregation - All tools from all servers available through one interface
  • - Authentication - One API key for all your MCP servers
  • - Analytics - See which tools are being used and how often

Learn More