What is MCP (Model Context Protocol)? A plain-English guide

What the Model Context Protocol is, how MCP clients and servers work, and why it matters for AI agents.

The short answer

The Model Context Protocol (MCP) is an open standard, introduced by Anthropic in November 2024, that lets AI applications connect to external tools and data in one consistent way. Instead of building a custom integration for every app, a developer builds one MCP server and any MCP-compatible client can use it.

Clients, servers and tools

An MCP client is the AI app you use, such as Claude Desktop, Cursor, VS Code or ChatGPT. An MCP server is a small program that exposes capabilities to that client. Servers offer three kinds of things:

• Tools: actions the model can call, such as running a SQL query or opening a web page.
• Resources: data the model can read, such as files or database rows.
• Prompts: reusable templates the user can pick.

How they talk

Local servers run on your computer and talk to the client over standard input/output (stdio). Remote servers run on the internet and use Streamable HTTP. In both cases the messages are JSON-RPC 2.0.

Why it matters

MCP turns a chat assistant into an agent that can act: read your repository, query a database, drive a browser or post to Slack. Because the protocol is shared, the same server works across many apps.

Where to start

Pick a server for a task you already do, install it with one of the guides below, and give it only the access it needs.