How apps talk to each other: APIs and MCP
An API explained with a restaurant waiter, and how MCP, an open standard, lets AI assistants use your apps with you in the loop.
Watch on LinkedIn (opens in a new tab)
Read the transcript
Every time your phone shows the weather, an app is asking another computer for help. That conversation runs through an API.
Think of a restaurant. You don't walk into the kitchen. You order from the menu, and the waiter takes it to the kitchen and brings back your dish. An API is the waiter. It carries your request to another system and brings back a response, and it only offers what's on the menu.
In an app, it looks like this: the weather app asks for Ottawa, and the weather service answers with data.
Now AI assistants want to use your apps too. But wiring every assistant to every app, one by one, becomes a tangle.
MCP, the Model Context Protocol, is an open standard for that. Its official guide compares it to a USB-C port for AI: one standard plug for any app that supports it.
Here's the conversation. The assistant asks the app: what can you do? The app replies with its tools. You say: book coffee with Sam on Tuesday. The assistant picks the right tool and asks before it acts. You approve, the app does the work, and the assistant confirms.
An API is the waiter between systems. MCP is the universal plug that lets AI use them, with you in the loop.
Azora. Good work, connected.
Sources
Checked on 18 October 2026. The restaurant is a teaching picture, not a definition.
- What is MCP? (opens in a new tab)Model Context Protocol
- Specification: Tools (opens in a new tab)Model Context Protocol
- Architecture overview (opens in a new tab)Model Context Protocol