Making the XML your business depends on visible.

XML never went away. It quietly runs sitemaps, RSS and Atom feeds, SOAP services, government open data, financial messaging, e-invoicing mandates, and a large share of business-to-business integration. Most of it is load-bearing and none of it is glamorous.

For the teams that depend on it, XML is usually a black box: parsed by a script someone wrote years ago, understood by one person who has since moved teams, and monitored not at all. The failure mode is always the same. A partner changes a field, a feed drops an element, a sitemap doubles in size — and you learn about it from the symptom rather than the cause.

XMLDir closes that gap. It turns scattered XML documents into an indexed, searchable, monitored system of record: extract a document's structure, validate it, compare it against the version you saw last week, and get told when the shape changes.

The goal is unremarkable and worth saying plainly — give the people who own these integrations the same confidence about their XML that they already expect from the rest of their stack.

Well-formedness and schema validation both run libxml2 — the same engine, compiled to WebAssembly — rather than a permissive JavaScript parser that accepts documents a real consumer would reject. That choice is the reason a verdict here matches the verdict your pipeline will give you.

The same engine formats documents, so indenting a file preserves mixed content, CDATA, comments, processing instructions and namespace prefixes exactly. Regex-based formatters quietly corrupt all five.

Every example on the format and errorpages is checked by the test suite against that parser on every build: the broken snippets are asserted to produce the exact message documented, and the fixed ones to parse cleanly. When libxml2's wording changes, the tests fail and the page gets corrected — rather than slowly becoming folklore.