Encoding Recovery
Repairing Corrupted Text
Drop a file or paste text, let it detect the encoding, and save it in the format you need.
Getting text in
Drop a file, or paste the garbled text directly. The encoding is guessed from the bytes as soon as it arrives, and shown at the top.
What it recognises
- BOM: the marker at the head of a file identifies UTF-8, UTF-16 LE, and UTF-16 BE outright.
- UTF-16 without a BOM: recognised by the dense pattern of null bytes — common in older SMI subtitles.
- UTF-8: every byte sequence is validated.
- CP949 / EUC-KR: chosen when the bytes decode to Korean syllables.
- Shift_JIS: chosen when Japanese kana appear.
Repairing mojibake
Press Repair for text like '¾È³çÇϼ¼¿ä' or '안녕하세ìš"'.
Corruption happens two ways: UTF-8 bytes read as Latin-1, and CP949 bytes read as Latin-1. Both reversals are attempted and the more plausible result wins — trying only one and accepting the first "improvement" turns the other kind into meaningless Korean.
Repair works when the characters changed but the information survived. Text already saved as question marks or empty boxes has lost the information and cannot be recovered.
Saving
- UTF-8: the default everywhere today.
- UTF-8 (BOM): use this if Korean breaks when Excel opens your CSV.
- EUC-KR: for older devices and programs that predate UTF-8.
- UTF-16 LE: required by some Windows programs.
Files are never uploaded. Reading, repairing, and saving all happen inside your browser.