Set your cleanup once.
Let new files clean themselves.

Yes. Save any cleaning sequence as a named recipe, then apply it in one click, run it across an entire folder on demand, or assign it to a watch folder. A watch folder processes every new or changed CSV dropped into it automatically, writing the cleaned result beside the original. Recipes can include formulas and prunes, so even complex workflows run hands-free. Built-in Recipe Packs give you a starting point for common sources without building from scratch.

Yes on both counts. CSVPrune runs entirely on your machine with no cloud, no telemetry, and no network access of any kind — there is no networking code in the app, so it works fully offline and nothing you load is ever uploaded. Your data stays in local memory for the session, and only minimal settings such as window state and recent paths (which you can turn off) are saved to disk. You can verify the zero-network behaviour with a connection monitor.

CSVPrune loads CSV files with comma, semicolon, or tab delimiters (auto-detected, with a manual delimiter and encoding override for UTF-8, UTF-16, and Latin-1) and is designed for large files, with smooth performance on hundred-thousand-row datasets and guidance up to around 500 MB. You can export your cleaned data to CSV, JSON, or XLSX, copy it to the clipboard, or use an audited export that adds a full before/after sidecar for compliance.

CSVPrune is a one-time purchase on the Microsoft Store — no subscription and no account required to use it. You own the version you buy and it keeps working offline. Use the Microsoft Store button on this page for the current listing and price, and because the privacy posture is structural rather than a setting, you can verify the local-only behaviour yourself at any time with a network monitor.