Mojibake Repair Tool
Paste text that came out as unreadable symbols and this tool works backwards through the encoding mix-ups that cause it, showing you the most likely original text. Japanese 文字化け, Korean encoding breakage and the classic European “Café” problem are all handled, entirely in your browser.
How mojibake happens, and how it is undone
Text is stored as bytes, and an encoding is the agreement about which bytes mean which characters. Mojibake happens when that agreement breaks: bytes written with one encoding are read back with another. Japanese written in UTF-8 and opened by a program expecting Shift_JIS turns into strings of unrelated kanji; the same text opened as Windows-1252 turns into strings of accented Latin letters.
The repair is the mirror image of the damage. Take the garbled text, encode it back into the encoding that was wrongly used to display it, and that recovers the original bytes. Decode those bytes with the encoding that was actually correct and the original text comes back. This tool runs that round trip across every common mismatch pair and ranks what comes out.
Ranking matters because the problem is genuinely ambiguous - the same bytes can decode into plausible-looking text in several encodings at once. Candidates are scored on how much they look like real writing, whether the round trip is self-consistent, and whether the writing system holds together (kanji mixed with kana reads as Japanese; kanji with no kana at all usually does not). That is why more than one answer is offered rather than a single confident guess.
One limit is worth being blunt about: if the text already contains replacement characters, those bytes were discarded before you ever copied it. No tool can bring them back. The text around them can still be recovered, but the gaps are permanent.
अक्सर पूछे जाने वाले प्रश्न
What does 文字化け (mojibake) mean?
Mojibake is the Japanese word for garbled text - literally 'character transformation'. It describes what you see when text is displayed using the wrong character encoding: Japanese turns into strings of unrelated kanji and half-width katakana, Korean turns into scrambled Hangul, and European text sprouts accented capitals like à and Â. The text itself is usually fine; only the interpretation of its bytes is wrong, which is why it can often be repaired.
文字化け 復元 ツール - can this restore my garbled text?
In most cases yes. Paste the broken text and the tool tries every common encoding mismatch, then shows the top candidates ranked by how much each result looks like genuine writing. Pick the one that reads correctly and copy or download it. The main case it cannot fix is text that already contains replacement characters, because those bytes were destroyed rather than merely misread.
Which encodings does it handle?
UTF-8, Shift_JIS (CP932), EUC-JP, EUC-KR (CP949), GBK (CP936), Big5, Windows-1252 and ISO-8859-1, in both directions. That covers the mismatches behind the overwhelming majority of real-world mojibake in Japanese, Korean, Chinese and Western European text, including legacy CJK files opened through a Western code page and text that went through the same broken conversion twice.
Why does it show several answers instead of just one?
Because mojibake repair is genuinely ambiguous. A single byte sequence can decode into perfectly valid-looking text in Shift_JIS, EUC-KR and GBK simultaneously - the tool cannot know which language you meant. Presenting a ranked shortlist and letting you recognise your own text is more honest, and more useful, than picking one and hiding the alternatives.
Why is some text still missing after repair?
If the garbled text contains the replacement character, the program that produced it hit bytes it could not interpret and threw them away. That loss happened before the text reached you and is permanent. The tool flags this when it detects it, recovers everything around the gaps, and marks the affected candidates so you know what to expect.
Is my text uploaded anywhere?
No. All the encoding and decoding runs locally in your browser using the platform's own text encoders. Nothing is sent to a server, which also means it works on confidential documents.
My CSV opens as garbage in Excel - is that the same problem?
Usually yes. Excel often assumes a legacy regional code page rather than UTF-8 when opening a CSV, which garbles any non-ASCII text. You can repair the text here, but the lasting fix is to import the file through Excel's Data > From Text/CSV option and explicitly choose UTF-8, or save the file with a byte order mark so Excel recognises the encoding by itself.