XLSX to CSV in the browser is not a generate
/free-tools/xlsx-to-csv never spends a credit. Do not pay a model to do an in-browser file job.
XLSX to CSV never spends a credit. Do not pay a model to do an in-browser file job. An .xlsx is a ZIP of XML parts. This tool reads xl/worksheets plus shared strings and writes RFC 4180 CSV on your device. Nothing is uploaded. Nothing is sent to a model. There is no account and no limit, because the work costs nothing to run.
Versely has no free tier for generation. Every generate spends credits. A "free converter" that quietly called a model would either contradict that or eat margin on anonymous traffic. This utility sidesteps both: it is free because it never touches the catalog.
What the parser actually emits
Pick a sheet, preview the rows, download. Shared strings, numbers, and dates are handled locally. Formulas emit the last cached value Excel stored in the cell — not the formula text. Macros are ignored. A .xls is the old binary format and is not a ZIP, so it cannot be read here. Re-save as .xlsx from Excel, Pages, or LibreOffice first.
Some dates come out as 44927. Excel stores dates as a serial day count. The tool converts built-in date number formats (14–22 and a few others). Custom formats it does not recognise are left as the serial. If a column of dates comes out as numbers, convert that column in a spreadsheet instead of asking a video model to "fix the CSV."
The file you choose is read into memory on your own machine, processed there, and offered back as a download. Disconnect the network and it still works. That is the privacy FAQ on every tool in this surface, and it is literal.
Credits are a different meter
What it costs is twelve billing shapes, not one price per file. A lipsync is metered by the second. A voiceover is per 1,000 characters. An image is a flat call or a megapixel. An edit is often the export. Credits are the common unit those meters resolve into, and the app quotes the figure before every charge. None of those formulas apply to unzipping XML in a tab.
The hub is explicit: no USD-per-unit figure, no provider cost table, customer-facing plan prices only on that surface. Spreadsheet-to-CSV is not on the list because it is not a billed job. Paying a text-to-image row 4 credits, or a video row dozens, to "turn this Excel into a table" is how you buy a picture of a table. You already had the table.
The free tools hub is the rest of the same rule: Canvas, the File API, or arithmetic, in the visitor's browser. Resize, subtitle sidecar, this CSV dump. If the next step is a generate — a chart as a motion plate, a voiceover of the numbers — that step is a catalog row, quoted in credits, after you already have a CSV you did not pay to parse.
Do not launder a file job through a prompt
If the sheet is the source of truth for SKUs, locales, or a caption list, parse it here. Then, if you must, feed rows to a model on purpose. Prompting "extract the spreadsheet" spends a generate to do a ZIP read, and you still will not get RFC 4180. Cached values, one sheet at a time, download. That is the whole product.
FAQ
Is there a catch — a watermark, an account, a daily cap?
No. Actually free, with no account and no limit, because it runs on your device rather than on Versely hardware. Generation is the thing that costs credits. This is not generation.
Why did my formulas become numbers?
Because Excel writes a cached value into the cell, and that is what the XML contains. The tool does not re-evaluate your workbook. If you needed the formula text, keep the .xlsx.
Can it read an old .xls?
No. .xls is not a ZIP of XML. Re-save as .xlsx and it becomes readable. Macros in a modern file are ignored either way.
Should I use a model when the dates come out as serials?
No. That is a custom date format the parser does not map. Convert the column in a spreadsheet. A generate will not know that 44927 was a Tuesday, and you will have spent credits on a wrong column.