Consuming External APIs (JSON & cURL)

Talking to the outside world

In the last unit you *served* JSON. Now you will *consume* it. Real apps constantly call other services — a weather API, a currency-rate API, a payment gateway. PHP can make an HTTP request to another server, receive its JSON, and turn it into PHP data. The standard tool for this is cURL, a library built into PHP for making web requests.

*Consuming an API* means your PHP acts like a browser: it sends a request to a URL and reads the response — except the response is JSON data, not an HTML page.

Fetching data with cURL

cURL follows a fixed four-step pattern: open a session, set options (the URL and "return the result to me"), run it, then close it.

<?php
  $url = "https://api.exchangerate.host/latest?base=USD";

  $ch = curl_init($url);                              // 1. open
  curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);     // 2. return result
  $response = curl_exec($ch);                          // 3. run
  curl_close($ch);                                     // 4. close

  $data = json_decode($response, true);   // JSON text -> PHP array
  echo "1 USD = " . $data["rates"]["INR"] . " INR";
?>

Step by step: curl_init($url) starts a request to the API. curl_setopt(..., CURLOPT_RETURNTRANSFER, true) is the crucial option — it tells cURL to return the response as a string instead of printing it. curl_exec($ch) actually sends the request and gives back the JSON text. After closing, json_decode($response, true) converts that JSON into a PHP associative array, so $data["rates"]["INR"] reads the exchange rate just like any nested array.

Handling failure

External calls can fail — the service might be down or return an error. Always check before trusting the data, so your page does not crash on a bad response.

<?php
  $response = curl_exec($ch);
  $httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);

  if ($response === false || $httpCode !== 200) {
    echo "Could not reach the service.";
  } else {
    $data = json_decode($response, true);
    // ...use $data...
  }
?>

curl_getinfo($ch, CURLINFO_HTTP_CODE) reports the status code the remote server sent. We treat the call as failed if curl_exec returned false (a network problem) or the code is not 200 (the server reported an error). Only when both checks pass do we decode and use the data — a small guard that keeps your page calm when the outside world misbehaves.