csvsort

Most CSV jobs are about shape, not cells: the columns are in the wrong order, the headers are someone else's names, half of them aren't needed, and the importer on the other end wants something else. csvsort does exactly that job. Drag columns into order (or move them with the keyboard), rename them, hide the ones you don't need, sort by any column, and filter rows by contains, equals, a numeric range or empty. What you see in the preview is what gets exported.

It reads what spreadsheets actually write: commas, semicolons from European Excel, tabs or pipes, detected from the first lines with a manual override; a leading UTF-8 BOM; files with or without a header row. Export picks the delimiter, LF or CRLF line endings, an optional BOM for Excel, and whether to include the header. A round trip is lossless, and a test pins that.

The app is a static React page and the file never leaves the tab. Large files parse in a Web Worker with a progress bar, and the preview scrolls through hundreds of thousands of rows. There is no server-side component, no account and no analytics.

What it is not: a spreadsheet. Cells are read-only, there are no joins or merges across files, and nothing is saved between sessions.