CSV to Excel
Convert CSV or TSV to XLSX while keeping text intact
CSV becomes XLSX text by default, preserving leading zeros, long numbers, dates and formula-like text. Optional typing accepts round-tripping numbers within 15 significant digits and lowercase true/false; dates remain text. Empty physical lines are skipped; missing fields become empty text and extra fields get column names. Limits: 20 MiB, 100,000 rows, 500 columns, 2,000,000 cells; 32,767 UTF-16 code units per cell.
Processed in your browser without saved drafts. Analytics excludes input, filenames, sheet names, column names and raw errors.
About this tool
Create a genuine XLSX from pasted CSV/TSV or a local file. All values remain text by default. Optional typing recognizes lowercase true/false and numbers that round-trip within 15 significant digits; dates, leading-zero identifiers and unsafe numbers stay text. Auto encoding uses a UTF-16 BOM or strict UTF-8, with explicit UTF-16 LE/BE, GB18030 and Windows-1252 alternatives. Invalid decoding fails instead of inserting replacement characters. Delimiter selection, generated or unique headers, extra-column preservation and a 30-row by 20-column preview help verify the result. Limits: 20 MiB, 100,000 rows, 500 columns, 2,000,000 cells including headers and 32,767 UTF-16 code units per cell.
Scenario Recipes
Create an XLSX while keeping account codes and date text intact
Goal: Control when plain CSV strings become typed spreadsheet values.
- Paste CSV/TSV or choose the actual file encoding before uploading. Auto uses a UTF-16 byte-order mark when present, otherwise strict UTF-8; failed decoding does not silently fall back to another charset.
- Choose the delimiter and whether the first row is a header. Leave type detection off to preserve all values as text. Optional detection recognizes only lowercase true/false and numbers that round-trip within 15 significant digits; leading-zero identifiers, unsafe numbers and dates remain text.
- Prepare the preview, inspect renamed columns and the detected delimiter, then download XLSX. Changing text or settings clears the preview; prepare again before downloading.
Result: A typed XLSX file whose text values retain identifiers, invalid date text and formula-like strings without formula calculation.
Failure Clinic (Common Pitfalls)
A leading zero disappears or an impossible date becomes a different day
Cause: Unrestricted numeric and JavaScript Date conversion changes plain CSV text.
Fix: Keep the default text mode. Optional typing does not recognize dates and retains noncanonical numbers such as 01, -001, +1 and numbers with edge spaces as text.
An automatic encoding guess produces readable but incorrect characters
Cause: Several encodings can decode the same bytes. Successful decoding is not proof of the original charset.
Fix: Auto is limited to BOM-based UTF-16 or strict UTF-8. Explicit UTF-16 LE/BE, GB18030 and Windows-1252 are available. Select the source encoding and upload again; invalid byte sequences produce an error.
A single-column file fails or a wider row appears to lose cells
Cause: A single column needs no delimiter, and data rows can be wider than the first header row.
Fix: Single-column CSV is accepted. Wider rows receive extra generated columns; shorter rows receive empty text. Empty physical lines are skipped, while quoted newlines stay within a field. Limits: 20 MiB, 100,000 rows, 500 columns, 2,000,000 cells and 32,767 UTF-16 code units per cell.
Production Snippets
Text defaults preserve values that look numeric or date-like
csv
id,long_id,date,formula
01,9007199254740993,2026-02-30,=1+2
Default XLSX cell types: text | text | text | text
The invalid date is not normalized to March.
Optional typing still leaves these four fields as text.Frequently Asked Questions
A leading zero disappears or an impossible date becomes a different day?
Unrestricted numeric and JavaScript Date conversion changes plain CSV text. Keep the default text mode. Optional typing does not recognize dates and retains noncanonical numbers such as 01, -001, +1 and numbers with edge spaces as text.
An automatic encoding guess produces readable but incorrect characters?
Several encodings can decode the same bytes. Successful decoding is not proof of the original charset. Auto is limited to BOM-based UTF-16 or strict UTF-8. Explicit UTF-16 LE/BE, GB18030 and Windows-1252 are available. Select the source encoding and upload again; invalid byte sequences produce an error.
A single-column file fails or a wider row appears to lose cells?
A single column needs no delimiter, and data rows can be wider than the first header row. Single-column CSV is accepted. Wider rows receive extra generated columns; shorter rows receive empty text. Empty physical lines are skipped, while quoted newlines stay within a field. Limits: 20 MiB, 100,000 rows, 500 columns, 2,000,000 cells and 32,767 UTF-16 code units per cell.
How do I use this converter?
Paste CSV/TSV or choose the actual file encoding before uploading. Auto uses a UTF-16 byte-order mark when present, otherwise strict UTF-8; failed decoding does not silently fall back to another charset. Choose the delimiter and whether the first row is a header. Leave type detection off to preserve all values as text. Optional detection recognizes only lowercase true/false and numbers that round-trip within 15 significant digits; leading-zero identifiers, unsafe numbers and dates remain text. Prepare the preview, inspect renamed columns and the detected delimiter, then download XLSX. Changing text or settings clears the preview; prepare again before downloading.
Is my input uploaded or saved as a draft?
Processing runs in this browser. Input is not uploaded or saved as a draft; analytics excludes values and filenames.
Keep browsing