Qu’est-ce qu’une API ?

An API (Application Programming Interface) is how one program talks to another. Instead of rebuilding maps, payments, or weather data yourself, you call someone else's service and get the answer back.

When you build software with an AI assistant, most real features come from connecting to third-party APIs. Learning to think in APIs lets you assemble powerful apps quickly.

Think of an API like a restaurant menu. You don't go into the kitchen and cook. You read the menu, place an order with the waiter, and food arrives.

The menu is the API documentation, your order is the request, and the dish is the response. You never see how the kitchen works inside, only what it promises to deliver.

Every API interaction has two halves: a request you send and a response you receive. The request says what you want; the response carries the data or a status.

You can ask your AI assistant to explain this flow for any service you plan to use, so you know exactly what to send and what comes back.

An endpoint is a specific URL where the API listens for one kind of request. A weather service might have one endpoint for current conditions and another for forecasts.

Each endpoint does a focused job. Knowing which endpoint to hit is half the battle when integrating any external service into your app.

APIs use HTTP methods to signal intent. GET fetches data, POST creates something, PUT or PATCH updates, and DELETE removes. The method tells the server what action you mean.

When unsure, ask your assistant which method an endpoint expects. Picking the wrong one is a common beginner mistake that returns confusing errors.

Most modern APIs speak JSON, a lightweight text format of keys and values. A weather response might include "temp": 21 and "city": "Paris".

JSON is easy for both humans and code to read. Your AI assistant can turn a raw JSON response into clean variables your app actually uses.

Many APIs require an API key, a unique string that identifies and authorizes you. It is like an ID badge that lets you through the door and tracks your usage.

Keys are secret. Never paste them into public code or share them in chat. We will cover safe handling in a later lesson.

APIs limit how many requests you can make in a time window, called a rate limit. Free tiers might allow 60 calls per minute, paid tiers far more.

Hitting the limit returns an error, not your data. Ask your assistant to check a service's limits before you build features that call it in a loop.

Every response carries a status code. 200 means success, 401 means unauthorized, 404 means not found, and 429 means too many requests.

These numbers are your fastest debugging clue. When something breaks, the status code usually points straight at the cause before you read any error text.

Some APIs are public and free, like many weather or quote services. Others are private, needing accounts, billing, or approval, like payment and messaging platforms.

Knowing which kind you face shapes your setup. Ask your assistant to compare options so you start with the simplest service that meets your need.

Building with AI is mostly about wiring services together. A trip planner might combine a maps API, a weather API, and a currency API into one screen.

Once you see features as integrations, you describe the outcome to your assistant and let it handle the plumbing of requests, parsing, and display.

Test your understanding of API basics.

An API lets your app borrow another service's powers through requests and responses. You hit endpoints with HTTP methods, send and receive JSON, and identify yourself with keys.

Status codes and rate limits guide your debugging. Next, you will learn to read API docs quickly with help from your AI assistant.