AI engineering · 12 of 42

Model Context Protocol

One standard interface for tools and context

Scroll

One standard interface for tools and context

Every tool an AI application can use — a repository, a document store, a database, a calendar — used to need its own integration, and every new client re-implemented all of them.

MCP defines one interface for discovering what a server offers, calling it, and getting a result back. Write a server once and any compatible client can use it.

What it standardizes is exposure, not permission. The protocol has nothing to say about who may call a tool, with whose credentials, or what gets audited. Those remain entirely your problem, and a standard interface makes them more urgent rather than less, because it is now much easier to connect things.

Interface
ONE INTERFACE INSTEAD OF ONE ADAPTER EACH your AI app (MCP client) requests MCP server one standard shape discover, call, return GitHub your docs a database calendar / files capabilities it exposes Without it, every tool needs its own bespoke integration and every new client re-implements all of them. The protocol standardises exposure, not permission. Who may call what, and with whose credentials, is still entirely your problem.
One app talking to many capabilities through a single standard interface rather than a bespoke adapter each.