What Is MCP · Model Context Protocol (MCP) in Generative AI | Generative AI & LLMs

The Problem

Every AI app (Claude Desktop, Claude Code, Cursor, VS Code, your own chatbot) wants to talk to the same systems: GitHub, Google Drive, Slack, databases, your company APIs. Without a standard, each app writes its own connector for each system: N apps × M systems integrations.

The Solution

The Model Context Protocol (MCP) is an open protocol, introduced by Anthropic in November 2024, for connecting AI applications to tools and data. You write one MCP server for your system, and any MCP-capable app can use it. People call it "USB-C for AI apps".

Without MCP                     With MCP
App A ─┬─ GitHub                 App A ─┐            ┌─ GitHub server
App A ─┼─ Postgres               App B ─┼── MCP ─────┼─ Postgres server
App B ─┼─ GitHub                 App C ─┘            └─ ShopEase server
App B ─┴─ Postgres
 (N × M connectors)               (N + M pieces)

MCP vs Tool Use

	Tool use (function calling)	MCP
What it is	an API feature: the model asks to call a function	a protocol for discovering and calling tools across processes
Where tools live	inside your app code	in a separate server, reusable by many apps
Relationship	the model still uses tool use underneath	the MCP client turns server tools into tool definitions

Note

Note: MCP does not replace tool use. The host app lists the MCP server's tools, passes them to the model as normal tool definitions, and forwards each tool call to the server.