An XML minifier removes the indentation and line breaks between tags, optional comments, and extra spaces inside tags, without touching text values, CDATA or quoted attributes. For example, an indented 19-byte document with root a and one child b holding 1 shrinks to 15 bytes, a 21.1% cut, with identical parsed content.
Pretty-printed XML is easy to read but carries a lot of dead weight: every level of nesting adds a newline and a run of spaces that a parser throws away. Developers trimming SOAP payloads, sitemaps, Android resource files or config bundles shipped inside an app usually want that weight gone. Paste or open a file and the minified XML appears straight away, along with the original and new size in UTF-8 bytes, the percentage saved, the element count and how many comments were removed. The minifier also checks that every opening tag has a matching closing tag and names the first one that does not. It is not a full validating parser: it does not check namespaces, entity definitions or a schema, and it assumes whitespace between elements is insignificant, which is wrong for mixed content such as XHTML paragraphs.
Paste the document into the box, or use Open a text file to load a .xml, .svg or .rss file from your device. The file is read locally.
Leave Remove XML comments ticked to drop every <!– –> block, or untick it to keep licence headers and notes.
Trim spaces around text values strips leading and trailing spaces in element text. Untick it when those spaces carry meaning.
Read the minified XML and the size figures, then copy the output into your file or request body.
Result Minified XML: <a><b>1</b></a>
Original size (bytes): 19
Minified size (bytes): 15
Bytes saved: 4
Size reduction: 21.1%
Elements: 2
Comments removed: 0
The newline and two-space indent around the child element disappear, leaving 15 of the original 19 bytes.
Result Minified XML: <note><to>Ana</to></note>
Original size (bytes): 46
Minified size (bytes): 25
Bytes saved: 21
Size reduction: 45.7%
Elements: 2
Comments removed: 1
The comment is dropped along with the indentation, and the two elements are counted.
Result Minified XML: <list><item id="1" name="a b"/></list>
Original size (bytes): 47
Minified size (bytes): 39
Bytes saved: 8
Size reduction: 17%
Elements: 2
Comments removed: 0
Runs of spaces between attributes shrink to one, but the double space inside the quoted value stays as written.
Saving % = (original bytes − minified bytes) ÷ original bytes × 100
original bytesminified bytesThe input is split into tags, comments, CDATA sections, processing instructions, declarations and text. Text runs made only of whitespace are dropped, following the XML 1.0 rule that such whitespace in element-only content is insignificant. Inside tags, runs of spaces and newlines collapse to a single space, but anything between quotes is copied unchanged. CDATA is never altered. Sizes are counted in UTF-8 bytes, so an accented letter counts as two bytes and most emoji as four.
For element-only content, no. A parser reports the same elements, attributes and text values; only the whitespace-only nodes between elements disappear. If your code reads those whitespace nodes, for example by counting child nodes rather than child elements, the count will change.
No. Everything between <![CDATA[ and ]]> is copied exactly, including line breaks and indentation, because CDATA content is character data the receiving application sees verbatim.
They are kept, with internal runs of whitespace collapsed to single spaces. The declaration must stay first in the document, so the minifier never moves or removes it.
While minifying, each opening tag is pushed onto a stack and each closing tag must match the most recent open one. If the closing tag for a arrives while b is still open, or an element is never closed, the tool stops and names the tag so you can fix the source first.
It depends on how deeply indented the document is and how long its values are. A heavily nested config with short values often shrinks by 20 to 40 percent, while a feed full of long descriptions may save under 10 percent. The size figures show the exact result for your input.
Yes, SVG is XML, so indentation and comments are removed the same way. Text inside SVG text elements is trimmed when that option is on, so untick it if leading spaces in a label matter.
They work together. Gzip already compresses repeated whitespace well, so minifying mostly helps where compression is not applied, such as embedded strings, database fields or size-limited message queues. Combining both gives the smallest transfer.