JSON Diff / Compare

JSON Formatter re-indents a JSON document so it can be read and reviewed, and points out the first syntax error if it is not valid. Paste minified or hand-broken JSON, choose 2 spaces, 4 spaces or a tab, and it is parsed by the browser's JSON engine and printed back cleanly. Parsing is strict — comments, trailing commas and single-quoted strings are rejected. Because it round-trips through the engine, duplicate object keys collapse to the last value, keys that look like integers are re-ordered numerically ahead of the rest, and integers past plus or minus 2^53 or very long decimals lose precision. For the opposite goal — the smallest payload — use JSON Minifier, which validates the same way but removes whitespace instead of adding it. Input is limited to 2 MB, about 100,000 values and 64 nesting levels, and nothing is uploaded.

JSON Minifier removes every space, tab and line break that JSON does not need, producing the most compact form of the same data. It validates first: an invalid document is rejected with a clear error rather than minified, so you never ship broken JSON. The values are untouched, but — like the formatter — a round-trip through the browser's JSON engine keeps only the last of any duplicate object key and can re-order integer-like keys. Use it to shrink a request body or an embedded config; use JSON Formatter for the reverse, when you need indentation to read or diff a document. There is no XML equivalent — the XML tool only re-indents. Input is limited to 2 MB, about 100,000 values and 64 nesting levels, and the document never leaves your browser.

XML Formatter re-indents an XML document and reports the first structural problem it finds. It checks well-formedness only — matched tags, exactly one root element, and no text outside it — and names the first fault (an unclosed tag, a mismatched closing tag, stray text, or more than one root). It does not validate against a DTD or XSD schema, resolve or check entities such as the ampersand escape, or verify namespace prefixes; an xmlns is treated as an ordinary attribute. Indentation is fixed at two spaces per level, and whitespace inside text nodes is trimmed, so significant spacing in mixed content is not kept. Unlike the JSON tools it never rebuilds the data model, so nothing is re-ordered; there is also no XML minifier. Input is limited to 2 MB, roughly 100,000 tags and 64 nesting levels, and the document stays in your browser.

Paste an XML document and get a deterministic JSON representation, produced entirely in your browser. Element attributes are kept under an "@attributes" object, and every attribute and text value stays a string, so "true" or "00123" is never turned into a boolean or a number. Repeated sibling elements become a JSON array while a single element stays a direct value, and the original element order is preserved rather than sorted. Namespace prefixes and xmlns declarations are kept as written, CDATA sections become plain text, and comments, processing instructions and the XML declaration are dropped. For safety the converter refuses any DOCTYPE or ENTITY declaration before parsing, so external entity references and entity-expansion attacks cannot be processed, and it rejects mixed text-and-element content, multiple root elements and malformed markup instead of guessing a correction. Input is limited to about 2 MB, conversion runs only when you select Convert, and nothing you paste is uploaded, logged or stored. JSON cannot represent every XML feature exactly, so unsupported structures are rejected or simplified according to the documented mapping.