Many Lights Tools

Fix garbled file names in ZIP files

Choose a zip and the tool shows its file names re-read in several character encodings. Once the names are readable, you can rebuild the zip with those names and save it. You can then unzip the new file the usual way. You can also extract files one at a time. Your zip never leaves your browser.

Made for people whose file names came out garbled when they opened a zip made on a Mac in Windows, or a zip made with old Windows software on a Mac.

or drop it here

One zip file. It is processed on this device only.

How to use

  1. Choose one zip file. You can also drop it onto the page.
  2. Look at the list of file names. The encoding that looks most readable is selected automatically at first. If the names are not readable, select another encoding.
  3. Press “Save a new zip with the fixed names”, then unzip the new file the way you usually do. To get just one file, press “Save” in the list.

Frequently asked questions

Is my zip uploaded anywhere?
No. Reading the zip, re-reading the names, rebuilding the zip and extracting files all happen in your browser. The page only sends anonymous counts: that a page was opened, and that an action such as "saved" happened. See the Privacy Policy.
Why do the names get garbled?
A zip can carry a flag saying that its file names are written in UTF-8. When a zip does not have this flag, the software that opens it decides which encoding to use, and the names come out garbled if that differs from the encoding used to write them. For example, older Japanese Windows software often writes names in Shift_JIS (CP932), and the built-in feature of a Mac may write names in UTF-8 without the flag.
How does the rebuilt zip differ from the original?
The file names are rewritten in UTF-8 and the UTF-8 flag is set. The contents are copied as the compressed data they already are (nothing is recompressed). Modification times are kept (to 2 seconds). The zip comment and the extra fields (such as finer timestamps) are not carried over. Symbolic links are left out, and file attributes are reset to those of an ordinary file or folder. Characters that cannot be used in file names (control characters, colons, question marks and so on) are replaced with _, and “../” and a leading “/” are removed. If two names end up the same (including names that differ only in upper and lower case), (2) is added to the later one. The original zip is not modified.
Is the automatically selected encoding always right?
No, it is a guess. Names that are valid UTF-8 are taken as UTF-8. Otherwise the tool selects a candidate that has the fewest unreadable characters and reads as letters such as kana and kanji (when candidates are close, the Japanese encodings Shift_JIS and EUC-JP are selected first). With short names, names made only of kanji, or Chinese, Korean or Western European names, the guess can be wrong. Look at the list and select another encoding if it is not readable. The candidates are UTF-8, Shift_JIS (CP932), EUC-JP, GB18030 (GBK), Big5, EUC-KR, CP437 (DOS), Windows-1252. Names that carry the UTF-8 flag are always read as UTF-8, whatever encoding you select.
Are there zips it cannot open?
Yes. Password-protected (encrypted) zips, ZIP64 (zips over 4 GB, or with 65,535 or more files) and split zips are not supported. Formats other than zip, such as 7z, rar and lzh, cannot be opened either. When this is the case, the tool says so.
Some files in the list cannot be saved with “Save”.
Files can be extracted one at a time if they are stored without compression or compressed with Deflate (the most common method in zips). Deflate is expanded with a feature of the browser. In a browser without that feature, only uncompressed files can be extracted. Files over 64 MB cannot be extracted one at a time either. In all of these cases, rebuilding the zip still works. A file extracted this way gets a single name in which folder separators are replaced with _.
What does “Join split dakuten marks into single characters” mean?
In file names made on a Mac, a character such as “が” may be stored as two characters: “か” and a separate dakuten (voiced sound) mark (a form called NFD). It looks almost the same, but Windows and other systems treat it as a different name when searching or sorting. With this option ticked, such characters are joined into one (NFC).
What are “__MACOSX” and “.DS_Store”?
They are system files that a Mac puts into a zip. They are not used outside a Mac. With the option ticked, they are left out of the rebuilt zip and are not shown in the list.
Does it cost anything?
It is free, and there is no sign-up.

Before you use it