JSON Diff & Compare

A semantic diff, not a text diff

A plain text diff reports a change whenever a key moves or the indentation shifts. This tool parses both documents first and compares the resulting values, so reordered object keys and reformatted whitespace are never reported as differences. Each side is then re-printed with matching indentation so the two panes line up row by row.

Arrays get special treatment. Instead of comparing index against index, identical elements are anchored with a longest-common-subsequence pass, so inserting one item at the top of a list marks a single insertion rather than rewriting every element below it. Elements that only partly match — same identifier, different values — are paired up and compared field by field.

Comparing JSON that is not valid yet

Real payloads rarely arrive clean. One side comes from a log where every quote is escaped as \"key\", the other was hand-edited and has single quotes, unquoted keys, and a trailing comma. With Auto-fix broken JSON enabled, each pane is repaired before the comparison runs and the repairs applied are listed under the editor, so you can see exactly what was altered. Press Fix & format both to write the cleaned, indented version back into the editors.

The repair pass handles, among others:

• Backslash-escaped quotes, newlines, and tabs from log output
• JSON that is double-encoded as a string inside another string
• Missing opening or closing braces on truncated lines
• Trailing commas, comments, single quotes, and unquoted keys
• Python and JavaScript literals such as None, True, and NaN

Controlling what counts as a difference

Two responses are often equivalent for your purposes even when they are not byte identical. The compare options let you say so:

• Ignore array order compares lists as unordered collections, so a reshuffled array of tags is identical.
• Loose types treats "42" and 42 as equal, which is useful when one service serializes numbers as strings.
• Ignore case and Trim strings normalize string values before comparing.
• Ignore keys drops noisy fields such as updatedAt or requestId from both sides so they never show up as changes.

Four ways to read the result

Side by side aligns both documents with insertions, deletions, and edits colour-coded, and highlights the changed part inside each edited line. Unified collapses that into a single column of + and - lines that pastes cleanly into a pull request or ticket. Changes lists every differing path with its old and new value, filterable, with one-click path copying. Patch emits an RFC 6902 JSON Patch that transforms the left document into the right one, ready to send to an API that accepts patch documents.

Turn on Only differences to collapse long identical stretches, and use the up and down arrows to jump between changes in large documents.

Why it is private

Parsing, repairing, and comparing all run as JavaScript inside this page. Open your browser DevTools, switch to the Network tab, and paste a document: no request carries the document. That matters when you are diffing API responses that contain tokens, customer records, or internal identifiers.