XML Tools

Indent 2, 4 or 8 spaces or tabs, wrap attributes when a start tag gets long, normalise empty elements and keep xml:space="preserve" subtrees exactly as they were.

Every structural error reported with its position and a plain-English reason: which element was left open, which closing tag was found, which attribute appears twice.

Type a path, see the matches as you type, and get a readable error when the engine hits an axis or function it does not implement rather than an empty result.

Point it at the repeating element and get a table: one row per record, one column per leaf, attributes marked with @ — ready for a spreadsheet.

Element order, occurrence, required attributes, unknown attributes and built-in types with facets — reported against the node that broke the rule.

Indent 2, 4 or 8 spaces or tabs, wrap attributes when a start tag gets long, normalise empty elements and keep xml:space="preserve" subtrees exactly as they were.

Shrink a document for transport or a log line, with a per-change report of exactly what was removed and how much it saved — and mixed content left intact.

The issues a plain validator cannot report because the document is still valid: mixed tabs and spaces, stale namespace declarations, a BOM that contradicts the encoding, repeated attribute values.

Take comments out of a document for a smaller payload or a cleaner diff, with a count of what was removed and a guarantee that comment-looking text inside CDATA survives.

Put markup, SQL or HTML inside XML without escaping every angle bracket, and split the terminator correctly when the text contains one.

Both directions of XML escaping, with the unknown entities reported instead of silently dropped — the difference between ` ` and a real non-breaking space.

Turn `café` back into `café`, drop the byte-order mark that breaks string comparison, and fix a declaration that says ISO-8859-1 while the bytes are UTF-8.

Line and column numbers for every row, so a message from a parser or a colleague ('line 214, column 9') points at the same place here.

See the shape of a document instead of its syntax: expand and collapse elements, read attributes at a glance, and copy the XPath that selects exactly the node you clicked.

Type a path, see the matches as you type, and get a readable error when the engine hits an axis or function it does not implement rather than an empty result.

The mapping a document actually declares, prefix by prefix, with the scope where each applies and the elements that use it.

List every `id`, `sku` or `href` in a document as a flat column — with duplicates highlighted and a count, ready to paste into a spreadsheet or a check script.

A quick profile of a document: how many of each element, how deep it nests, which names dominate, and where the bytes actually go.

A diff that understands documents: inserted records are reported as additions rather than shifting every following line, and a reformatted file is not reported as a rewrite.

Point it at the repeating element and get a table: one row per record, one column per leaf, attributes marked with @ — ready for a spreadsheet.

A readable YAML rendering of a document, using the @attribute convention and text keys, with types coerced only when every value agrees.

Concatenate exports into one document — wrapper root or the first document's root reused — with a `source` attribute recorded on each merged part so the origin is never a guess.

Turn a batch export into one file per record, each still a valid document with its ancestors rebuilt around it and the file name taken from a key attribute.

Write or paste a stylesheet, watch the transform run as you type, and see the result rendered plus the exact serialized output.

Every structural error reported with its position and a plain-English reason: which element was left open, which closing tag was found, which attribute appears twice.

Element order, occurrence, required attributes, unknown attributes and built-in types with facets — reported against the node that broke the rule.

Give it one or two real documents and get a starting-point schema: types guessed from the values, minOccurs and maxOccurs from what repeats, required attributes from what always appears.

Content models, attribute types, required and fixed values, ID uniqueness and IDREF resolution — validated against the DTD as a real parser would.

The mistakes that make a feed silently fail in readers: a missing channel link, duplicate GUIDs, an Atom-style date in an RSS pubDate, entries with no alternate link.

See what a document asks a parser to do before you parse it: external and parameter entities, an external DTD subset, XInclude, and the entity-expansion patterns behind billion-laughs.

Build a request skeleton with header and body parameters, correct namespace prefixes, escaped values and the right transport header line for your SOAP version.

Indent 2, 4 or 8 spaces or tabs, wrap attributes when a start tag gets long, normalise empty elements and keep xml:space="preserve" subtrees exactly as they were.

Every structural error reported with its position and a plain-English reason: which element was left open, which closing tag was found, which attribute appears twice.

Shrink a document for transport or a log line, with a per-change report of exactly what was removed and how much it saved — and mixed content left intact.

The issues a plain validator cannot report because the document is still valid: mixed tabs and spaces, stale namespace declarations, a BOM that contradicts the encoding, repeated attribute values.

Take comments out of a document for a smaller payload or a cleaner diff, with a count of what was removed and a guarantee that comment-looking text inside CDATA survives.

Put markup, SQL or HTML inside XML without escaping every angle bracket, and split the terminator correctly when the text contains one.

Both directions of XML escaping, with the unknown entities reported instead of silently dropped — the difference between ` ` and a real non-breaking space.

Turn `café` back into `café`, drop the byte-order mark that breaks string comparison, and fix a declaration that says ISO-8859-1 while the bytes are UTF-8.

Line and column numbers for every row, so a message from a parser or a colleague ('line 214, column 9') points at the same place here.

See the shape of a document instead of its syntax: expand and collapse elements, read attributes at a glance, and copy the XPath that selects exactly the node you clicked.

Type a path, see the matches as you type, and get a readable error when the engine hits an axis or function it does not implement rather than an empty result.

The mapping a document actually declares, prefix by prefix, with the scope where each applies and the elements that use it.

List every `id`, `sku` or `href` in a document as a flat column — with duplicates highlighted and a count, ready to paste into a spreadsheet or a check script.

A quick profile of a document: how many of each element, how deep it nests, which names dominate, and where the bytes actually go.

A diff that understands documents: inserted records are reported as additions rather than shifting every following line, and a reformatted file is not reported as a rewrite.

Point it at the repeating element and get a table: one row per record, one column per leaf, attributes marked with @ — ready for a spreadsheet.

A readable YAML rendering of a document, using the @attribute convention and text keys, with types coerced only when every value agrees.

Concatenate exports into one document — wrapper root or the first document's root reused — with a `source` attribute recorded on each merged part so the origin is never a guess.

Turn a batch export into one file per record, each still a valid document with its ancestors rebuilt around it and the file name taken from a key attribute.

Write or paste a stylesheet, watch the transform run as you type, and see the result rendered plus the exact serialized output.

Element order, occurrence, required attributes, unknown attributes and built-in types with facets — reported against the node that broke the rule.

Give it one or two real documents and get a starting-point schema: types guessed from the values, minOccurs and maxOccurs from what repeats, required attributes from what always appears.

Content models, attribute types, required and fixed values, ID uniqueness and IDREF resolution — validated against the DTD as a real parser would.

The mistakes that make a feed silently fail in readers: a missing channel link, duplicate GUIDs, an Atom-style date in an RSS pubDate, entries with no alternate link.

See what a document asks a parser to do before you parse it: external and parameter entities, an external DTD subset, XInclude, and the entity-expansion patterns behind billion-laughs.

Build a request skeleton with header and body parameters, correct namespace prefixes, escaped values and the right transport header line for your SOAP version.