Fix the weird characters in a CSV

Fixing the characters in a CSV file

1. Drop the file in. It is read by JavaScript in this tab. No server sees it, and closing the tab throws it away.
2. Read what it detected. Above the table it says "Read as …". If the accents and names look right, that was the one.
3. Export as UTF-8. Download .csv or .xlsx — the output is always UTF-8, so it does not break again downstream.

Why “José” becomes “José”

A CSV does not record the encoding it was written in. It is a run of bytes, and whoever opens it has to guess. In UTF-8 an é takes two bytes; read those two bytes as Windows-1252 and you get two characters, Ã and ©. Hence the unmistakable pattern: Ã followed by something odd.

The other direction gives black diamonds with question marks (�): the file was Windows-1252, Shift-JIS, Big5 or EUC-KR, and something is reading it as UTF-8.

How the detection works here

1. If there is a BOM, the encoding is declared and there is nothing to guess.
2. Without one, UTF-8 is tried in strict mode. Most files that are not UTF-8 fail that, which already rules a lot out.
3. Among the survivors the decoded text is scored and the best one wins. Scoring counts runs of CJK characters, not isolated ones, and counts accented Latin letters as evidence too — that distinction was added on 2026-09-04 after a Spanish Windows-1252 file was being decoded as GBK, because a page of Western high bytes read as Chinese scores well under a naive "more CJK is better" rule.
4. Windows-1252 is a full candidate, not a last resort.

You are the real check: if the names read correctly, it picked the right one.

Stopping Excel from breaking it again

On Windows, double-clicking a UTF-8 CSV with no BOM makes Excel read it in the system's regional encoding — and it breaks all over again. Two ways out:

• Download .xlsx instead of .csv. A workbook carries its encoding inside it.
• If it has to be CSV, open it through Data → Get Data → From Text/CSV and choose 65001 (UTF-8) in the dialog.

All in one column is a different problem

If the file opens with everything crammed into column A, that is the delimiter, not the encoding: in many European locales Excel expects a semicolon because the comma is the decimal separator. CSV to Excel detects the delimiter on its own.