Compare Two CSV Files

If any column has a unique value in every row of both files (an ID column is the common case), it’s auto-detected and used to match rows by VALUE instead of position - so a row that moved is still recognized as the same row, not reported as removed-and-re-added. Without a unique column, rows are compared by position using the same alignment algorithm behind line-by-line text diffs: it correctly finds the longest run of untouched rows and reports only what actually changed around them, but a fully shuffled set of otherwise-identical rows (with no unique column) will show as removed and re-added, the same limitation ordinary text diff tools have.

With a key column, this is exact: the same key means the same row, so any cell difference between the two is a "changed" row. Without a key column (comparing by position), a deleted row directly next to an inserted row is reported as one "changed" row only if enough of its cells still match - otherwise they’re unrelated content that happened to land next to each other, and showing them as separate removed/added rows is the more honest read.

No. Both files are read and compared entirely on your device. Turn off your Wi-Fi after the page loads and it still works. This page loads 18KB of JavaScript, gzipped - about what your browser's network tab will show for this page's own scripts.

Each file is capped at a generous size (shown on the drop zone) so a single huge file can’t freeze the tab. Separately, comparing by row position uses an algorithm whose work grows with both files’ row counts multiplied together, so extremely large files (roughly two files each in the low thousands of rows or more) may be refused for position-based comparison with a message pointing you at the key-column option instead - matching by a unique column has no such limit.

Every row from the comparison (unchanged rows included, regardless of the on-page "show unchanged rows" toggle) with a "Status" column (Unchanged/Changed/Added/Removed) and, for any cell that changed, both values shown as "old → new" in that cell.

Correctly - this tool reads both files with a full CSV parser, so a quoted field like "Smith, John", or one containing its own line break, is read as one field, not split apart.

Case-sensitive by default ("Rent" and "rent" count as different) and whitespace-sensitive by default (a leading/trailing space counts as a change) - both are toggles above the table if you want the comparison to ignore either.

A one-time $3 unlock (no subscription) for two things: batch mode, comparing many file pairs in one pass instead of one at a time, and a combined Excel report - one workbook, one sheet per pair, with changed/added/removed rows highlighted. Everything above stays free and unaffected either way.