乱码修复

UpBox优盒工具箱，免费在线乱码修复：把「UTF-8 被当 Latin-1 显示」这类乱码贴进来，自动尝试两种反向解码并按中文占比给出可信度，日志、接口返回、CSV 里的乱码都能试一把。

无需注册 · 数据不上传No sign-up · nothing uploaded

乱码原文Garbled text

已识别

最常见的一类是 UTF-8 字节被当 Latin-1 或 Windows-1252 显示，于是中文变成「Ã¤Â½」这种组合；另一种是文件本身是 Latin-1 却被按 UTF-8 读。工具会两种都试，按中文占比排出可信度

可能的结果Candidates

只处理「字节顺序没错、只是解码时用错了编码」这一类；如果是文件传输被截断、编码表本身就是 GBK 与 Big5 错配，本工具修不回来，需要按原始来源重新导出

本工具所有运算在本地浏览器完成，你的文件、文本不会上传到服务器

工具介绍

日志里看到「Ã¤Â½」，接口返回一串「ÃÂ」，第一反应是数据坏了；其实字节完好，只是解码时选错了编码表。 Seeing "Ã¤Â½" in a log or a wall of "ÃÂ" from an API looks like corrupted data, but the bytes are usually intact and only the wrong table was used to decode them.

把这段乱码贴过来，工具会正反两个方向各试一次，谁读起来更像正常中文就排在前面。 Paste the garbled text and the tool tries both directions, putting the reading that looks most like real text first.

双向尝试Both directions
可信度排序Ranked by plausibility
中文占比CJK share
一键复制Copy result

使用说明

贴乱码。Paste the garbled text.
看候选结果。Read the candidates.
选最像的一条。Pick the best one.
复制回去替换。Copy it back.

常见问题

为什么会出现这种乱码？Why does this happen?

文件是 UTF-8 编码，读取的程序却按 Latin-1 或 Windows-1252 解释，一组三字节的中文就变成了三个怪符号。The file is UTF-8 but the reader interprets it as Latin-1 or Windows-1252, so one three-byte Chinese character becomes three odd symbols.

中文占比是怎么算的？How is the CJK share computed?

统计结果里中日韩字符和全角标点占全部字符的比例，比例越高越像正确还原。It counts CJK characters and full-width punctuation as a share of all characters; the higher the share, the more likely a correct restore.

GBK 乱码能修吗？Can it fix GBK mojibake?

部分可以：如果 GBK 文本被当 Latin-1 存过，会出现「ÁË」这类双字节组合，本工具能试出来；纯粹的 Big5 与 GBK 错配则修不了。Partly: GBK text saved as Latin-1 shows "ÁË"-style pairs which this tool can recover, but a pure Big5-versus-GBK mismatch cannot be fixed here.

文件被截断的乱码呢？What about truncated files?

修不了。字节缺失后无法反向推出原字符，只能回到原始来源重新导出。Not recoverable: missing bytes cannot be inverted, so export the original again.