Skip to content
LanguageHelper

Réparateur d’encodage cassé (mojibake)

Retransformez « Café crème » en « Café crème » et « Привет » en « Привет ».

Options

0

Résultat
Télécharger

 

100 % dans votre navigateur

À propos de cet outil

Mojibake appears when text saved in one character encoding is read with another — most often UTF-8 bytes decoded as Windows-1252, which turns “é” into “é” and Cyrillic into “Ð…Ñ…” sequences. Because no information is lost, the damage can be reversed. This tool reconstructs the original bytes and re-decodes only valid UTF-8 sequences, so correctly encoded characters in mixed text are left untouched, and it repeats the process for double or triple encoding. In auto mode it also tests legacy encodings — Windows-1251, KOI8-R, Shift-JIS and GBK — and picks the most plausible result. Auto mode tries 19 legacy encodings — Cyrillic (Windows-1251, KOI8-R/U, CP866), Central European (Windows-1250), Turkish, Greek, Hebrew, Arabic (Windows-1256), Baltic, Vietnamese (Windows-1258), Thai (TIS-620), Japanese (Shift-JIS, EUC-JP), Chinese (GBK, Big5) and Korean (EUC-KR) — and keeps the result that reads most like real language.

Comme tous les outils de LanguageHelper, Réparateur d’encodage (mojibake) fonctionne entièrement dans votre navigateur. Votre texte n’est jamais envoyé, stocké ni utilisé pour l’entraînement, ce qui le rend sûr pour les documents confidentiels, les données clients et les messages personnels. Aucun compte ni installation n’est nécessaire, et l’outil continue de fonctionner hors ligne une fois chargé.

Comment ça marche

  1. Paste the garbled text.
  2. Keep “Auto-detect” or choose what happened if you know.
  3. The repaired text appears instantly.
  4. Copy it, and re-save the source file as UTF-8 to prevent it happening again.

Cas d’usage

  • Repair CSV files opened with the wrong encoding in Excel.
  • Fix garbled names and addresses in databases and CRM exports.
  • Recover old e-mails, subtitles and web pages with broken characters.
  • Clean scraped text before analysis or translation.

Exemples

UTF-8 as Windows-1252

Avant

Café crème

Après

Café crème
Cyrillic

Avant

Привет

Après

Привет

Conseils pour de meilleurs résultats

Copy the garbled text exactly as it appears; don’t retype it. If auto mode doesn’t fully fix it, try the specific encodings — Russian files often used Windows-1251 or KOI8-R and Japanese files Shift-JIS. After repairing, save your files as UTF-8 and make sure the program reading them is set to UTF-8 too.

Questions fréquentes

What causes mojibake?

A program reads bytes with the wrong encoding: a CSV exported as UTF-8 opened in Excel as Windows-1252, a database connection with the wrong charset, or an e-mail missing its charset header.

Can every garbled text be fixed?

Text is recoverable when the wrong decoding kept all bytes, which is the usual case. If characters were replaced with “?” or “�” during saving, the original information is gone and cannot be restored.

What is double encoding?

Text that was mis-decoded and re-saved twice, producing sequences like “é”. The fixer repeats the repair up to three times until the text is clean.

Will it damage text that is already correct?

No. Only byte sequences that form valid UTF-8 are re-decoded, and correctly encoded characters such as “naïve” or “—” are preserved.

Which encodings are supported?

UTF-8 misread as Windows-1252 or ISO-8859-1 (including double encoding), plus 19 legacy encodings covering all 20 site languages: Windows-1250, 1251, 1253, 1254, 1255, 1256, 1257, 1258, KOI8-R, KOI8-U, CP866, ISO-8859-2, TIS-620/Windows-874, Shift-JIS, EUC-JP, GBK/GB18030, Big5 and EUC-KR. Choose one explicitly if auto-detection picks the wrong one.

How can I prevent mojibake in the future?

Save files as UTF-8, declare the charset in HTML and e-mail headers, set database connections to utf8mb4, and when opening CSV files in Excel use Data → From Text/CSV with UTF-8 selected.