How to Use the Mojibake Repair Guide
If you've ever opened an English document and seen quotation marks or dashes replaced with odd character clusters, you've run into mojibake β usually caused by a UTF-8 file being opened as Windows-1252 (or "ANSI"), or vice versa. Since curly quotes, em dashes, and symbols like the trademark sign occupy the 0x80-0x9F byte range in Windows-1252, a multi-byte UTF-8 sequence for one of those characters gets split into several separate, wrong characters when read under the other encoding.
Just paste the garbled text into the box above. The tool checks for lost characters (replacement characters, which mean the original bytes are already gone and unrecoverable) and, if none are found, attempts to reverse the Windows-1252 byte mapping and re-decode the result as UTF-8. If that produces a different, cleaner string, it's shown as the repair attempt along with a plain-language diagnosis of what likely happened.
This approach recovers common cases β broken smart quotes, dashes, ellipses, and trademark symbols β but it can't undo damage from files that were re-saved multiple times under mismatched encodings, since those bytes are permanently lost.
Frequently Asked Questions
No β if the original bytes were already lost through multiple re-saves (shown as replacement characters), no amount of re-interpreting the encoding can recover the original text. In that case, re-obtain the original file and re-save it directly as UTF-8.
Curly quotes, dashes, and other special characters live in the 0x80-0x9F byte range in Windows-1252. When a UTF-8-encoded file, which uses multi-byte sequences for those characters, gets opened as Windows-1252 instead, each byte is shown as its own separate character, producing garbled multi-character sequences.