XML to JSON Converter

How to use the XML to JSON Converter

1. Paste XML on the left, or open or drop an .xml file. The document is checked for well-formedness first.
2. Read the JSON on the right. Each element becomes a key, attributes get a prefix, and text that shares an element with them sits under "#text".
3. Decide whether attributes are kept, whether numbers and booleans are typed, and whether repeated tags are always lists.
4. Copy the JSON or save it as data.json.

Examples

<?xml version="1.0" encoding="UTF-8"?>
<catalog updated="2024-05-01">
  <item sku="LMP-204" featured="true">
    <name>Desk lamp</name>
    <price currency="EUR">39.90</price>
    <tags><tag>office</tag><tag>led</tag></tags>
  </item>
  <item sku="RUG-017">
    <name>Wool rug</name>
    <price currency="EUR">129</price>
    <tags><tag>floor</tag></tags>
    <note/>
  </item>
</catalog>

<library>
  <shelf id="A">
    <book>Tidal Maps</book>
    <book>Lantern Hours</book>
  </shelf>
  <shelf id="B">
    <book>Quiet Engines</book>
  </shelf>
</library>

<rss version="2.0">
  <channel>
    <title>Harbour Notes &amp; News</title>
    <item>
      <title>First post</title>
      <link>https://example.test/p/1</link>
      <description><![CDATA[<p>Hello <b>world</b></p>]]></description>
    </item>
  </channel>
</rss>

About this tool

XML and JSON model data differently, so every converter has to make choices: XML has attributes, mixed text and tags that repeat, while JSON has only objects, arrays and plain values. This tool makes those choices explicit. Elements become object keys, attributes become keys with a prefix such as @id, text that sits beside attributes or child elements is stored under "#text", and an empty element becomes an empty string.

Repeated tags are the classic trap. A naive conversion turns two <item> tags into an array but a single <item> into an object, so code that reads the result breaks on the first short document. With "Always use arrays for repeated tags" switched on, every tag that appears more than once anywhere in the document is written as a list wherever it appears, giving consumers one predictable shape. Type conversion is deliberately careful too: 42 and 3.5 become numbers and true and false become booleans, but 007, phone numbers and 20-digit identifiers stay strings.

Before anything is converted the document is checked: an unclosed tag, a bare & or a second root element is reported with its line and column rather than producing half a result. Comments, processing instructions and the XML declaration are not data and are left out. Tip: the JSON Tree Viewer on this site is a comfortable way to explore a large result.