1. Home
  2. Text
  3. Mojibake Fixer

Mojibake Fixer

Paste garbled text and see the original, with the encoding mix-up reversed.

For every format and option in one place, use Encoder and Decoder →

Runs in your browser · No upload
Samples
Most likely fix | UTF-8 text that was read as Windows-1252
It’s a café – résumé

Your text is processed in your browser only and never sent to a server.

How to use

  1. Paste the garbled text into the box above. A sample is already filled in, so you can see a result right away.
  2. In Find automatically, the “Most likely fix” appears at the top. Candidates are ranked by how much they look like real writing.
  3. The label next to each candidate says what went wrong, for example “UTF-8 text that was read as Windows-1252”. Copy the result with the button on its top right.
  4. If you want to test one specific pair, switch to Choose manually and pick the encoding it was wrongly read as and the encoding it really was.

Why text turns into garbage

Text is stored as bytes. To show it, software has to guess which encoding those bytes use. If the file was saved as UTF-8 but the reader assumes Windows-1252, you get strings like ’ and é. Nothing was destroyed, the same bytes were just read with the wrong table. To fix it, run the wrong table backwards to recover the bytes, then read them with the right table. That is exactly what this tool does.

Original Garbled What happened
It’s café It’s café UTF-8 bytes read as Windows-1252
내용 ë‚´ìš© UTF-8 bytes read as Windows-1252
한글 ÇÑ±Û EUC-KR (CP949) bytes read as Latin-1
こんにちは 縺薙s縺ォ縺。縺ッ UTF-8 bytes read as Shift_JIS

The classic signs are ’ for a curly apostrophe, – for an en dash, é for é and ü for ü. These usually come from a UTF-8 page served without a charset, a database connection using latin1, or a CSV opened in Excel. The values in the table were produced by this tool’s own conversion functions.

When it cannot be fixed

  • Text that contains �: when UTF-8 Korean is read as EUC-KR, 한글 becomes something like �湲. Bytes that could not be interpreted were replaced by �, so the information is gone and the text alone cannot be restored. Go back to the original file and open it with the correct encoding. For whole files, Text Encoding Converter detects the encoding for you.
  • Question marks: if a program already replaced characters with ? when saving, the originals cannot be known.
  • Characters lost while copying: Windows-1252 has five unassigned bytes (0x81, 0x8D, 0x8F, 0x90, 0x9D). If the invisible control characters they produce get dropped on copy, the text may no longer be reversible.
  • Very short input: one or two characters give the ranking little to work with, so paste a full sentence when you can.

Double mojibake

If garbled text is saved as UTF-8 again and then misread a second time, it breaks another layer, for instance 안녕 becomes 안녕. Automatic mode tries up to two rounds, but only when both rounds are UTF-8 misread as a single-byte encoding.

FAQ

Why does my text show ’ instead of an apostrophe?

The text is UTF-8, but it was displayed as Windows-1252. The curly apostrophe is three bytes in UTF-8, and each of those bytes was drawn as its own Windows-1252 character. Paste it here and it turns back into ’.

Why do I see é instead of é?

Same cause: é is two bytes in UTF-8 (C3 A9), shown as the two characters à and ©. Fix the page’s charset to UTF-8, or paste the text here to repair what you already have.

Can question marks or black diamonds be fixed?

Not from the text alone. ? and � mean the original characters were already discarded. Reopen the source file or email with the correct encoding instead.

Which encodings does it try?

It assumes the text was read as Windows-1252, Latin-1, EUC-KR, Shift_JIS, GBK or Big5, and the real encoding was UTF-8, EUC-KR, Shift_JIS, GBK or Big5. Every pair that decodes cleanly is scored and sorted by how natural the result looks.

Is my text uploaded?

No. Everything runs in your browser and nothing is sent to a server.

Can't find the tool you need?

Tell us what you want to do. We build the most requested tools first.