This is not surprising, but was unknown to me: There is a DOM parser available in JavaScript, that you can use to parse an HTML string, and act on the resulting HTMLDocument object.

Is behaves roughly like the global document, but without the location stuff. The important part: you can do query selectors or child collections:

const parser = new DOMParser();
const htmlString = '<div class="title"><p class="date"><time datetime="2024-12-07T23:00:00.000Z">Dec 8, 2024</time></p><h1>DOM Parser API</h1></div>';
const parsedHtml = parser.parseFromString(htmlString, 'text/html');

parsedHtml.childNodes[0].childElementCount; // -> 2
parsedHtml.querySelector('.date').childNodes[0].getAttribute('datetime'); // -> "2024-12-07T23:00:00.000Z"

A newer alternative is the static method Document.parseHTMLUnsafe(input).

Source: https://developer.mozilla.org/en-US/docs/Web/API/DOMParser/parseFromString