A product CSV from a supplier almost never matches what your store wants. Column names differ, prices carry a currency symbol, some rows have no SKU, and the importer rejects the whole file over one bad cell. Catalog CSV reads the file in the page, maps the columns you have to the columns the importer expects, and checks every row before you export.
The file is read with the browser File API and stays on your machine. There is no upload step and no copy kept on a server.
The export is plain CSV with the headers you mapped to, so it loads into Shopify, WooCommerce, Square, Etsy, and most marketplace importers. You pick the target header names during mapping.
No. The CSV is parsed by JavaScript in the open tab. Close the tab and nothing is left, apart from the column mapping saved in your own browser.
Tens of thousands of rows work on a normal laptop. Very large feeds are slower because the whole file is held in memory to check it.
Filter the rows you want, export, then filter again for the next batch. Each export is a separate download.