Chinese Mojibake Fixer (乱码修复)
Paste garbled Chinese text — like ä¸æ–‡ or 锟斤拷 — and this tool automatically tries the most common encoding mix-ups (UTF-8, Latin-1, Windows-1252, GBK, Big5) to recover your original text. No signup, works in your browser.
Fix your garbled text
Why Chinese text turns into "ä¸æ–‡"
Garbled Chinese text almost always comes from the same root cause: text encoded in one character set gets decoded by something that assumes a different one. The most common version is UTF-8 text — which uses multiple bytes per Chinese character — getting wrongly read one byte at a time as Latin-1 or Windows-1252, the encodings used for Western European text. Each original byte becomes its own separate character, and when that garbled sequence gets saved or transmitted as UTF-8 again, you end up with strings like ä¸æ–‡ instead of 中文.
Because no actual data was thrown away in this process — every original byte is still in there, just mislabeled — it's fully reversible. This tool automates the byte-level detective work: it tries reversing the most common mix-ups (UTF-8, Latin-1, Windows-1252, GBK/GB18030, and Big5) and shows you every plausible result, ranked by confidence, so you can pick the one that actually reads correctly.
When mojibake can't be fixed
Not all garbled text is recoverable. If invalid byte sequences were replaced with a placeholder character — like the famous 锟斤拷 pattern common in Chinese developer circles — the original bytes are already gone. Multiple different original characters can collapse into the same placeholder, so there's no way to tell which one was there. If our tool detects this pattern, it says so plainly rather than guessing at a "fix" that would just be more garbage. The only real solution at that point is to go back to the original source (the database, the file, the API response) and re-export or re-read it with the correct encoding declared from the start.
Common encodings at a glance
| Encoding | What it's for |
|---|---|
| UTF-8 | The modern universal standard — used by virtually all websites, APIs, and modern software |
| GBK / GB18030 | Legacy and government-mandated encodings for Simplified Chinese, still common in older Mainland China systems |
| Big5 | Legacy encoding for Traditional Chinese, historically used in Taiwan and Hong Kong |
| Latin-1 / Windows-1252 | Western European single-byte encodings — not designed for Chinese at all, which is exactly why misusing them on Chinese text produces such recognizable garbage |
If encoding mismatches keep showing up in your content pipeline — CMS exports, database migrations, translation vendor deliverables — that's usually a sign of a deeper technical setup issue rather than a one-off accident. Our Baidu SEO and Chinese AI-SEO/GEO teams can review your technical stack as part of a broader Chinese market readiness audit.
Related tools
Frequently Asked Questions
-
Yes. The tool is completely free to use with no sign-up, registration, or account required. Paste garbled Chinese text and it automatically tries the most common encoding mix-ups to recover the original.
-
This happens when text encoded in one character encoding, usually UTF-8, gets decoded by software that wrongly assumes a different encoding, usually Latin-1 or Windows-1252, and then the garbled result gets saved or displayed as UTF-8 again. This double round-trip is reversible because no information is actually lost, just mislabeled at one step.
-
Usually not, and the tool will tell you why. This specific pattern typically appears when Chinese text was decoded with the wrong encoding and invalid byte sequences were replaced with a placeholder character before the text reached you, which permanently destroys the original bytes. Unlike the reversible cases this tool fixes, there is no byte-level trick that recovers text that has already been replaced. The only real fix is going back to the original source and re-exporting or re-reading it with the correct encoding.
-
It tries several common encoding mix-ups (UTF-8, Latin-1, Windows-1252, GBK/GB18030, and Big5) and checks two things for each: whether converting back and forth reproduces the exact same bytes with no data loss, and what percentage of the resulting text is made up of real Chinese characters. Results that pass both checks are marked high confidence. You can review every candidate the tool tried, not just its top guess, since the correct fix is easy for a human eye to confirm.
-
Your text is sent to our server for analysis but is not logged, stored, or shared. After the response is returned, the input is discarded.