Is my input uploaded?

No. Everything stays in your browser. encoding-japanese (MIT) and the built-in Latin-1 codec run entirely in client-side JavaScript.

Which mojibake patterns does it cover?

Eight (source, misread) pairs: (UTF-8, Shift_JIS), (UTF-8, EUC-JP), (UTF-8, Latin-1), (Shift_JIS, UTF-8), (Shift_JIS, EUC-JP), (EUC-JP, UTF-8), (EUC-JP, Shift_JIS), and (Latin-1, UTF-8). That covers almost every mojibake you'll meet day-to-day with Japanese text.

Can I recover text that's already filled with U+FFFD?

If the file was actually saved with U+FFFD replacement characters, the original bytes are gone and cannot be recovered. If you still have the file, reopen it in your editor with the correct source encoding (e.g. Shift_JIS) and paste the result here.

How is the Japanese-likeness score computed?

We add the ratio of (hiragana + katakana + CJK kanji) characters, then subtract the ratio of U+FFFD and C0 control characters (newlines and tabs excluded). Candidates scoring zero or below are filtered out and the rest are sorted descending.

I have whole CSV files to re-encode

Use the CSV Encoding Convert tool for batch file conversions. This tool is focused on recovering text that you already have on hand as a string.

Back to Japanese

Mojibake Fix

Paste garbled Japanese text and recover the original. Tries every (source → misread) encoding pair (Shift_JIS / EUC-JP / UTF-8 / Latin-1), ranks candidates by Japanese-likeness score. Runs entirely in your browser.

japanesetextconversion

How to use

Paste your garbled text and click Recover. The tool tries every plausible (source, misread) encoding pair across Shift_JIS / EUC-JP / UTF-8 / Latin-1 (ISO-8859-1), scores each candidate by Japanese-likeness (ratio of hiragana / katakana / kanji minus ratio of U+FFFD and control chars), and lists the top candidates ranked. Candidates identical to the input or filled with U+FFFD are dropped automatically. Nothing leaves your browser — conversions run through encoding-japanese (MIT) and a built-in Latin-1 codec.

FAQ

Is my input uploaded?: No. Everything stays in your browser. encoding-japanese (MIT) and the built-in Latin-1 codec run entirely in client-side JavaScript.
Which mojibake patterns does it cover?: Eight (source, misread) pairs: (UTF-8, Shift_JIS), (UTF-8, EUC-JP), (UTF-8, Latin-1), (Shift_JIS, UTF-8), (Shift_JIS, EUC-JP), (EUC-JP, UTF-8), (EUC-JP, Shift_JIS), and (Latin-1, UTF-8). That covers almost every mojibake you'll meet day-to-day with Japanese text.
Can I recover text that's already filled with U+FFFD?: If the file was actually saved with U+FFFD replacement characters, the original bytes are gone and cannot be recovered. If you still have the file, reopen it in your editor with the correct source encoding (e.g. Shift_JIS) and paste the result here.
How is the Japanese-likeness score computed?: We add the ratio of (hiragana + katakana + CJK kanji) characters, then subtract the ratio of U+FFFD and C0 control characters (newlines and tabs excluded). Candidates scoring zero or below are filtered out and the rest are sorted descending.
I have whole CSV files to re-encode: Use the CSV Encoding Convert tool for batch file conversions. This tool is focused on recovering text that you already have on hand as a string.

Related tools

CSV / text encoding converter — Shift_JIS ↔ UTF-8 / BOM / newlines

Re-encode CSV and text files between Shift_JIS (CP932), UTF-8, UTF-16LE and EUC-JP — fix Excel's mojibake on UTF-8, hand UTF-8 text to legacy systems that need Shift_JIS, or add BOM so Excel reads UTF-8 correctly. Add / remove BOM, swap newlines (CRLF / LF / CR), and auto-detect the input encoding. Batch convert and grab the result as a ZIP. Files never leave your device — everything runs in the browser.

developertextconversion

Hiragana ⇄ Katakana converter — bulk character mapping

Convert between hiragana and katakana with a single mode toggle. A purely mechanical per-character mapping — no dictionary download, instant conversion. Long-vowel mark, punctuation, kanji, and alphanumerics are preserved as-is. Runs entirely in your browser.

japaneseconversion

Fullwidth ⇄ Halfwidth converter — alphanumerics, kana, symbols

Convert between fullwidth ASCII (letters, digits, symbols) and halfwidth ASCII with a single mode toggle. Covers U+FF01–U+FF5E ⇄ U+0021–U+007E, plus the ideographic space U+3000 ⇄ ASCII space U+0020. Hiragana, katakana, and kanji are kept untouched. Runs entirely in your browser.

japaneseconversion

Kanji Numeral Converter

Convert between kanji numerals and Arabic digits. Supports place-value form (千二百三十四), daiji (壱弐参拾 — the tamper-resistant form used on contracts and promissory notes), and positional form (二〇二六). Handles up to 垓 (10^20), recognizes daiji, full-width digits, and thousands commas. Runs entirely in your browser — no uploads.

japaneseconversiontext