Kim Yihyung
← Tools
0 bytes uploaded · 0 external requests In-browser 100% local JSON ↔ CSV · format · minify · sort keys line or structure diff

See what changed.
Without the upload.

Paste two versions of a document: compare text line by line, or let both sides be parsed and compare JSON keys and array positions. Added, removed and changed rows are counted and coloured, and either side can be formatted, minified, key-sorted or converted between JSON and CSV. Nothing is uploaded and no library comes from a CDN.

Indent
Left · original —
Right · changed —

Result—

added0 removed0 changed0 unchanged0

Two documents, one comparison. Paste something into both panes, or press a JSON or CSV example button above.

Nothing leaves the tab

Both panes are ordinary text areas; parsing and comparison run as JavaScript in this page. The footer counts every request made outside this site, and it stays at zero — an API response or a customer export stays where it is.

Lines or structure

The line diff compares the text as written, so re-indenting shows as change. The structure diff parses both sides first and compares keys and array positions, so it finds that $.orders[2].total changed even after a reformat.

What to do with the result

Copy the report into a ticket, or download it as .json for a reviewer or .csv for a spreadsheet. Each pane also keeps its own copy and download, in the format it currently holds.

Questions

Can it compare files instead of pasted text?

Paste the file contents — open it in any editor and copy. There is no file picker here on purpose: a pasted pane leaves nothing on disk, and a diff rarely needs more than the part that changed.

How large a document can it handle?

Line comparison is quick when the two sides are similar, because unchanged head and tail are trimmed first. Beyond roughly 2 500 changed lines at once the page falls back to pairing lines by position, which is fast but coarser; JSON structure comparison is limited only by memory.

Why is my CSV followed by a JSON array of objects?

The first row is read as column names and every later row becomes an object with those keys. Values stay strings so leading zeros and long IDs survive; tick “Typed values in CSV” to turn numbers, true/false and null into their JSON types.

Note. A formatting difference is usually not a data difference: if the structure diff reports nothing, the two documents hold the same values even when they do not look the same. Check the CSV conversion before pasting it into a spreadsheet with leading-zero IDs.

No third-party library: the diff, the JSON formatter and the CSV reader and writer are part of this page. Everything is served from this site with the page, never from a CDN.