Model Context Protocol

What is Model Context Protocol (MCP)?

Definition: The Model Context Protocol (MCP) is an open standard from Anthropic that lets AI applications connect to external tools and data through a common interface. It's often described as "USB-C for AI" — one protocol, many integrations.

MCP standardizes how an AI client (Claude, Cursor, an agent framework) talks to an MCP server that exposes tools, resources and prompts. Messages use JSON-RPC 2.0 over a transport (today streamable HTTP for remote servers, or stdio locally).

The ecosystem passed 5,800+ community servers in the official registry, with Google and Microsoft shipping their own agent registries — making MCP one of the fastest-growing developer ecosystems of the AI era.

What does MCP actually standardize?

The handshake and message format between an AI client and a tool server: initialize, capability negotiation, and how tools/resources/prompts are listed and called — so any compliant client can use any compliant server.

Put it into practice

Health-check any MCP server free →

Related terms