Your own tags.
Strict rules.
Still everywhere.

02 · The building blocks

🧱 Elements, attributes & namespaces

An XML document is a tree. Here's a small one with every common part labelled in the comments:

<?xml version="1.0" encoding="UTF-8"?>          <!-- declaration (prolog) -->
<catalogue xmlns="https://example.com/ns/shop">   <!-- root element + default namespace -->
  <product id="TEA-250" currency="INR">   <!-- element with attributes -->
    <name>Assam Tea 250g</name>             <!-- element with text content -->
    <price>249.00</price>
    <tags>
      <tag>gift</tag>
      <tag>organic</tag>
    </tags>
    <discontinued/>                      <!-- self-closing empty element -->
  </product>
</catalogue>

🌐 Namespaces in one paragraph

Because anyone can invent tag names, two vocabularies can collide — your <title> and someone else's. A namespace ties names to a URI so they stay distinct. xmlns="…" sets the default namespace for an element and its children; xmlns:img="…" declares a prefix you then write as <img:image>. The URI is just an identifier — nothing is ever downloaded from it.

03 · Two levels of correct

✅ Well-formed vs valid

XML has two distinct bars, and people mix them up constantly.

A formatter or parser tells you whether a document is well-formed. Checking validity needs the schema too. Paste a document into the XML Formatter and it will point at the first well-formedness error.

04 · Special characters

🔣 Entities, escaping & CDATA

Because < starts a tag and & starts an entity, those characters can't appear raw in text. XML predefines exactly five entities:

Any Unicode character can also be written as a numeric reference: é or é for é. Note that HTML's named entities such as   and © are not defined in XML unless a DTD declares them.

<!-- escaped -->
<query>price &lt; 500 &amp;&amp; stock &gt; 0</query>

<!-- CDATA: everything up to ]]> is literal text -->
<query><![CDATA[price < 500 && stock > 0]]></query>

05 · Where you'll meet it

🗺️ XML vs JSON, and where XML still wins

For a fresh web API, JSON is the sensible default — see What is JSON?. But you'll keep running into XML here:

A sitemap is a good real example of a namespaced document:

<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
  <url>
    <loc>https://example.com/</loc>
    <lastmod>2026-03-14</lastmod>
  </url>
</urlset>

Want to check a real site's sitemap — every URL, its status code and whether it resolves? The Sitemap Checker finds and parses it for you.

06 · Querying

🎯 XPath basics

XPath is a small path language for selecting nodes in an XML tree. It's built into most XML libraries, browser DevTools and XSLT. Against the catalogue.xml example above (ignoring its namespace for a moment):

07 · In your code

🛠️ Parsing XML safely

Each snippet reads the sitemap above — namespace included, because that's what real documents look like.

using System.Xml.Linq;

XNamespace sm = "http://www.sitemaps.org/schemas/sitemap/0.9";
var doc = XDocument.Parse(xml);

var urls = doc.Descendants(sm + "url")
              .Select(u => (string?)u.Element(sm + "loc"))
              .ToList();

import xml.etree.ElementTree as ET

ns = {"sm": "http://www.sitemaps.org/schemas/sitemap/0.9"}
root = ET.fromstring(xml_text)
urls = [loc.text for loc in root.findall("sm:url/sm:loc", ns)]

const doc = new DOMParser().parseFromString(text, "application/xml");

// DOMParser doesn't throw — it returns a document containing <parsererror>
if (doc.querySelector("parsererror")) throw new Error("Not well-formed");

const urls = [...doc.getElementsByTagNameNS(
  "http://www.sitemaps.org/schemas/sitemap/0.9", "loc"
)].map(el => el.textContent);

Got a wall of minified XML from an API log? The XML Formatter indents it, checks it's well-formed and lets you explore it as a tree. If the same data also comes as JSON, compare it in the JSON Formatter.