Excel to CSV - Online Tool

1. Convert a workbook in three steps

The UI intentionally stays simple: pick a file, get CSV. If your workbook has multiple sheets, export only covers the first sheet SheetJS resolves. To convert another sheet, isolate it in Excel (copy to a one-sheet workbook) or use a sibling Excel tool that exposes a sheet picker, then come back here.

1. Choose a local .xlsx or .xls file. The tool reads it as an ArrayBuffer in your tab — there is no upload progress bar because nothing leaves the device.
2. Wait for the CSV text to appear in the output panel. Spot-check the first few and last few rows: header names, empty trailing columns, and any cell that contained commas or quotes.
3. Copy to the clipboard or download export.csv (named from the source file). Open the result in a text editor before feeding it to an importer so you catch encoding and quoting issues early.

2. What CSV keeps — and what it drops

CSV is a rectangular text grid. Cell values survive; cell styles, conditional formatting, charts, comments, pivot caches, and embedded images do not. Merged cells flatten to whatever SheetJS emits for the top-left (or equivalent) value — do not expect merged layout to round-trip. Formulas export as calculated values already stored in the file; if someone never recalculated before saving, you may see stale numbers. Treat the output as a snapshot of the grid, not an equivalent of the workbook.

• One sheet per CSV file. Multi-sheet books need one export each.
• Long numeric IDs and phone numbers often become scientific notation when re-opened in Excel — import those columns as text in the next tool.
• Dates may appear as serial numbers or locale strings depending on how the sheet stored them; verify with a known date row before batch work.

3. Quoting, commas, and newlines (RFC 4180 habits)

SheetJS sheet_to_csv wraps fields that contain commas, quotes, or newlines and doubles internal quotes — the usual RFC 4180 pattern. That is enough for most importers, but not every downstream parser is strict. Failure sample: a cell with a literal newline in an address field can look like two rows if a consumer ignores quotes. Failure sample: a field that already contains a double quote must appear doubled inside the quoted field; if a hand-edited CSV breaks that rule, re-export from this page instead of patching by eye.

• Prefer downloading the file over pasting through chat apps that reflow lines.
• If a database loader rejects a row, open the CSV in a plain-text editor and jump to that line — check for an unclosed quote earlier in the file.
• Delimiter is the SheetJS default comma. Semicolon-heavy European Excel locales still produce comma CSV here; convert separators only if your importer demands it.

4. When CSV is the right hop — and when it is not

Use CSV when a script, SQL loader, or Git-friendly diff needs plain text; when a partner system only accepts CSV; or when you want to strip macros and objects and keep the grid. Skip CSV when you need multiple sheets in one artifact, when formulas must remain formulas, or when nested JSON is the real target. Native JSON exports from the source system usually beat Excel→CSV→JSON guessing. Sibling pages cover Excel→JSON, Excel→Markdown, and CSV→SQL when those shapes fit better.

5. Privacy and handover notes

Spreadsheets often hold phone numbers, orders, and payroll. Local parsing avoids random online converters, but downloads still land in your Downloads folder and clipboard history can retain rows. Clear temporary copies on shared machines. When you hand a CSV to someone else, add one line of context: which workbook and sheet it came from, that only the first sheet was exported, and any columns that are text-not-number. That note prevents the classic “why did my employee IDs lose leading zeros” thread.

6. Acceptance checklist before you trust the file

1. Row count roughly matches the sheet (header plus data rows).
2. A known messy cell (comma inside text, multiline note) still parses as one field in your target importer.
3. Identifier columns still look like identifiers after the first import attempt.
4. You kept the original .xlsx until the importer succeeds — CSV is lossy relative to the workbook.