Compare XML without letting formatting hide the real changes

Direct answer

Does Diff Anything compare XML as text or structure?

Diff Anything compares parsed XML structure so indentation or other presentation-only changes do not overwhelm the result. Repeated elements are matched by a key attribute such as id or name, and each difference is located with a readable XPath. Pro can retain review state and notes in saved profiles and export the result without uploading the documents.

What is compared

Elements, attributes, and text, without layout noise.

Pretty-printing an XML file or reordering an element's attributes changes many lines of text but none of the content. Diff Anything reads both documents as elements, attributes and text, so indentation between elements and attribute order do not count as changes.

Text inside each element is trimmed, and whitespace-only text between elements is ignored. CDATA sections are compared as part of the element's text. Namespaces are compared by URI, so a document that uses a different prefix for the same namespace is not reported as changed.

Each difference is listed with a readable XPath location that uses the document's own prefixes, for example /catalog/@version for an attribute of the root or /catalog/m:item[@id="1"]/text() for the text of a keyed element.

Matching elements

Repeated elements are matched by name and key attribute.

Child elements are matched by their name and the first key attribute they have: id, key, name, code, ref or uuid, as in item[@id="1"]. Children without a key attribute are paired in order by name. An inserted or removed child is reported as added or removed, so the siblings after it are not shifted. When keyed children only moved, the result reports that the child elements were reordered.

If child order does not matter for a particular element, open Unordered elements and add its path as the document writes it, for example /root/items. Its children are then paired regardless of order, so a different order is not a change. This applies only to elements with no mixed text content.

XML opens in the Tree view, which shows both documents side by side. The Semantic tab lists every difference, and the Raw tab shows the two documents as written.

Boundaries

Comments and custom entities are outside the comparison.

XML comments, processing instructions, and the DOCTYPE are not compared. The predefined entities (&, <, >, ", '), numeric character references, and entities declared in the document's DOCTYPE are expanded, in text and attribute values. Undeclared, parameter and external entities, and entities whose text contains markup, stop the comparison with an unsupported-entity error. Nothing is fetched from outside the file.

Copying a change into an XML file rewrites the document from its elements. The confirmation says so before anything changes: the declaration, anything around the root element, and the document's own namespace prefixes are kept, but comments inside the root and indentation are not.

SVG files open in XML mode, so you can compare the markup of two graphics. To compare how two images look, use the Binary / Image mode with PNG or JPEG exports.

Size and plans

How large a document it handles, and what Pro adds.

One comparison reads documents of up to 500,000 elements, nested up to 256 levels. A larger or deeper document is reported as an error instead of being compared in part.

The free plan shows the complete XML result in the app for files up to 5 MB each, with 10 comparisons per day. Pro removes the 5 MB limit and adds saved profiles that keep your Unordered elements paths, comparison history, and HTML, JSON, or Markdown reports.

Best for

Where this workflow fits.

• Large configuration or data documents with formatting churn
• Private XML files that cannot be pasted into an online tool
• Repeatable reviews that need retained sources, mode, review states, and notes

No accounts, cloud comparison service, analytics, or telemetry.

Relevant capabilities

What Diff Anything brings to the job.

Practical workflow

From setup to a reviewed result.

1. Open both XML documents

   Choose the files or paste the markup. Auto opens XML and SVG files in XML mode.

2. Review the result

   Start in the Tree view to see both documents side by side, or open the Semantic tab for the list of changed elements, attributes and text.

3. Handle order-free lists

   Add paths under Unordered elements for lists where sibling order has no meaning.

4. Follow each location

   Use the XPath location of each difference to find the element, attribute or text that changed.

Useful in practice

Direct answers

Questions about XML diff tool.

Go deeper

Choose sources, understand automatic detection, and select the engine that answers the question you are actually asking.