CSVFixLab
Encoding

Fix CSV encoding and convert to UTF-8

Strictly confirm UTF-8. If the file is not UTF-8, choose the known source encoding before a UTF-8 repair is allowed.

Fix CSV encoding and convert to UTF-8
Auto confirms UTF-8 exactly; it does not pretend to identify an unknown legacy encoding.
Local browser processing
Drop a CSV file hereor choose a file from your device
Auto mode confirms UTF-8 exactly. If the file is not UTF-8, choose its source encoding before downloading a repair.
Analyzing…
—

CSV report

Health score reflects checks on this page; it does not guarantee import acceptance.

—data rows
—columns
—delimiter
—encoding
Download is enabled only when the parser can safely re-serialize the file without guessing the source encoding or missing business data.

Why encoding errors happen

CSV is a text format, but the bytes still need an encoding. A file written as Windows-1252, EUC-KR, or Shift-JIS can look correct in the program that created it and become garbled when another system assumes UTF-8. Names such as José, Müller, or Korean and Japanese text are often where the problem first becomes visible.

Why Auto mode stops at “Non-UTF-8 ?”

The browser can validate UTF-8 strictly, but it does not expose a dependable universal charset detector for arbitrary legacy CSV files. CSVFixLab therefore makes a deliberate distinction: it can prove that a byte sequence is valid UTF-8, but it does not claim that a non-UTF-8 file must be Windows-1252, EUC-KR, or Shift-JIS.

When Auto cannot validate UTF-8, the repair button remains disabled until you select the source encoding you actually know.

A reproducible example

This fixture contains accented characters encoded as Windows-1252. Auto can prove that the bytes are not valid UTF-8, but the user must confirm the source encoding.

Problem CSV
name,city,note
José,Montréal,Café résumé
Müller,Zürich,Crème brûlée
Structurally corrected CSV
name,city,note
José,Montréal,Café résumé
Müller,Zürich,Crème brûlée

What the checker can prove from the file

  • The visible characters are the intended data; only the byte encoding changes during a safe conversion.
  • Auto should report an unknown non-UTF-8 source instead of inventing a charset label.
  • After Windows-1252 is explicitly selected, the same text can be decoded and re-exported as UTF-8.
Safe next action

Confirm the source encoding from the system that produced the file, then convert and re-open the UTF-8 copy to verify non-ASCII text.

Do not guess

Do not repeatedly open and save a garbled preview; that can turn a reversible encoding mismatch into permanent character corruption.

Reproduce the behavior

Windows-1252 → UTF-8 lab caseUse the same fixture that exercises the Auto confirmation safeguard.

Read the underlying rule

CSV encoding and UTF-8 →Export CSV safely from spreadsheets →

How to verify the converted file

  1. Use Auto first. If the file is valid UTF-8, no conversion is needed for encoding alone.
  2. If Auto says Non-UTF-8 ?, identify the source encoding from the export settings or originating application.
  3. Select that source encoding and review the preview/report.
  4. Download the repaired copy only when the CSV structure is otherwise safe.
  5. Upload the repaired file again with Auto. A successful conversion should now validate as UTF-8.
Important: choosing the wrong source encoding can produce plausible-looking but incorrect text. CSVFixLab cannot recover the original bytes after another program has already replaced characters with question marks or replacement symbols.