Processed entirely on your device — nothing is uploaded
How to use the excel to csv
- 1Drop your .xlsx workbook into the box above.
- 2Pick the sheet you want, if the workbook has more than one — a CSV holds exactly one.
- 3Choose the delimiter and line endings, and add a byte-order mark if the file is going back into Excel on Windows.
- 4Press Convert to CSV and download the file.
The quoting rules that most converters get wrong
CSV looks trivial and is not. The format is specified by RFC 4180, and the rules that matter are about escaping: a field containing the delimiter, a double quote or a newline must be wrapped in double quotes, and a double quote inside such a field must be doubled. Get that wrong and the damage is silent — a description field containing a comma splits into two columns, and every column after it in that row shifts. Nobody notices until the import into some other system produces nonsense a week later.
This tool applies those rules on the way out and the same parser this site's CSV to JSON tool uses on the way in. A cell containing `Smith, John` is written as `"Smith, John"`. A cell containing a quotation is written with its quotes doubled. A cell containing a line break — which Excel allows, with Alt+Enter — is quoted so the row survives as one row.
The delimiter is your choice, and it is not always a comma. Semicolons are the default in much of Europe, where the comma is the decimal separator; tabs give you a .tsv, which sidesteps the whole quoting question for most data; pipes are common in data engineering. The tool quotes correctly for whichever you pick.
Dates, leading zeros and the two encoding traps
Excel stores a date as a number of days since 30 December 1899 and records that it is a date only in the cell's number format, in a different part of the file. Converters that ignore the format table turn every date into a five-digit number — 45292 instead of 2024-01-01. This one reads the formats, recognises both the built-in date formats and custom ones, and writes real dates in ISO order, which sorts correctly as text and is unambiguous about which number is the month.
The second trap is character encoding, and it produces the most common complaint about CSV files generated anywhere. The file here is written as UTF-8, which is correct and what every modern tool expects. Excel on Windows, however, only reads a CSV as UTF-8 if the file starts with a byte-order mark; without one it falls back to a legacy code page and accented characters, currency symbols and non-Latin text appear as mojibake. So there is a switch for the BOM. Turn it on if the file is going back into Excel; leave it off if it is going into a database, a script or almost anything else, since some parsers treat the mark as part of the first column heading.
Line endings are the third choice. RFC 4180 specifies CRLF, and strict Windows tooling wants it; everything else is happier with LF. Both are offered rather than assumed.
One sheet per file, and what happens to formulas
A CSV file holds a single table, so a workbook with several sheets converts one sheet at a time. The tool lists every sheet with its row count and lets you pick, appending the sheet name to the downloaded filename so a set of exports stays distinguishable.
Formulas are exported as their last calculated value, because that is what the file stores — a spreadsheet caches the result of every formula alongside its definition. If a workbook has never been opened and recalculated since a change, the cached values are what you get, which is the same behaviour as any other export.
Everything else that makes a spreadsheet a spreadsheet is discarded, as it must be: formatting, colours, merged cells, charts, images, comments, filters and multiple sheets have no representation in a plain text table. What you are left with is the data, which is exactly what CSV is for and why it remains the most portable format there is.
Why do this locally
The spreadsheets people convert to CSV are, characteristically, the ones full of things worth protecting: customer lists, price books, payroll runs, transaction exports, medical rosters. The standard online converter uploads that file to a server, converts it there and gives you a link — which is a copy of your data on infrastructure you do not control, under a retention policy you have not read.
Here the workbook is unzipped and parsed in this browser tab and the CSV is assembled in memory. There is no upload, no queue and no size limit beyond your own device's memory. A hundred-thousand-row sheet converts in a moment because there is no network round trip at either end.
It also works with no internet at all once the page has loaded, which is genuinely useful on a locked-down machine — the sort of environment where installing software is not an option and uploading a customer list would be a disciplinary matter.
Frequently asked questions
Is my workbook uploaded?
No. The .xlsx is unzipped and read inside this browser tab, and the CSV is assembled in memory. Nothing is transmitted — which matters, because the spreadsheets people export are usually customer, payroll or transaction data.
Can it convert every sheet at once?
A CSV file holds exactly one table, so sheets are converted one at a time. Pick the sheet from the list and repeat for the others; each download is named after its sheet.
Why are my dates coming out as numbers elsewhere?
Because Excel stores dates as days since 30 December 1899 and flags them as dates only in the number format. This tool reads that format table and writes real ISO dates instead.
My accented characters look wrong in Excel. What do I do?
Switch on the byte-order mark option. Excel on Windows only recognises a UTF-8 CSV as UTF-8 when the file starts with one; without it, it falls back to a legacy code page.
Can it read an old .xls file?
No — only .xlsx and .xlsm. The old binary .xls is a different format entirely. Open it in Excel or LibreOffice and save as .xlsx first.
What happens to formulas?
They are exported as their last calculated value, which is what the workbook stores alongside each formula. Formatting, charts, colours and merged cells are dropped, as they have no representation in CSV.