API stands for Application Programming Interface. Which tells you absolutely nothing. So let's ignore the name and talk about what it actually does.
You're at a restaurant. You know what you want. The kitchen can make it. But you don't walk into the kitchen yourself — you tell the waiter, the waiter takes your order back to the kitchen, and comes back with your food.
That's an API. It's the waiter.
(the app)
(the waiter)
(the kitchen)
When two apps talk to each other, they do it through an API. When you connect Claude to Notion, or your email, or your calendar — that connection happens through an API. You're essentially giving Claude a menu of what the service can do, and Claude places the orders.
This matters for a practical reason: as you start using AI for more things, you'll keep hearing the word "API." It usually just means the bridge between two tools. An API key is a unique string of letters and numbers that identifies you — like a member card for the restaurant.
This website actually has its own tiny API. When you click Subscribe and type in your email, the form sends a small request across to a piece of code that lives on the same site — and that code catches your email and saves it. There's no API key in the mix this time, because keys are mostly only needed when one site needs to talk to a different site. But the same kind of conversation is happening behind the scenes. You never see it. It just works.
Most APIs for AI tools have a free tier — meaning you can connect things and use them without paying, up to a certain amount. Which is a very good reason to start experimenting.