A formatter, not a parser
The XML کوڈ فارمیٹر splits a document at the boundaries between tags and re-indents it, adding two spaces on entering an element and removing them on leaving — which turns a document that arrived as one unbroken line, the way XML usually comes back from an API or an export, into something with a readable tree shape. What it does not do is validate anything along the way: it re-indents structure it recognises rather than checking that structure is actually correct.
Drifting indentation is often the first real warning sign
A document with a missing or mismatched closing tag is re-indented just as willingly as a correct one, with no error reported anywhere — the fault shows up only as indentation that drifts in a direction it should not, rather than returning to where it started once the relevant closing tag appears. That drift is frequently the first visible clue that something upstream produced a malformed document.
Mixed content is where it is least reliable
The split works at tag boundaries, which handles pure element nesting cleanly but is less predictable for mixed content — ordinary text with inline elements sitting inside it. That combination is the case most likely to come out looking different from what a person would have chosen by hand.
Attributes are left exactly as they were
Order, spacing and quoting inside a tag are untouched throughout — only the indentation between tags changes. Anything inside the tag itself is not this tool’s concern at all.
Use it to read, not to feed back in
Reformatted output is meant for looking at. Whitespace can be semantically significant depending on the schema a document is validated against, so whether reformatted XML is safe to send back into whatever system produced it is a question this tool has no way to answer on its own behalf.
Where it partly works on HTML, and where it does not
HTML permits tags that never get explicitly closed, which XML’s grammar does not allow, so running HTML through this produces the same kind of indentation drift a genuinely malformed XML document would show — not because the HTML is wrong, but because this formatter is reading it as something stricter than it actually is. A formatter built specifically for HTML is the better tool for that job.