Skip to content
ZeroServer.tools

ZIP File Extractor

Open a ZIP to see what it holds, view and save its files, or extract them all to a folder. Everything happens in this tab; nothing is uploaded.

ZIP file

Drop a ZIP here, or click to choose oneAlso .docx, .xlsx, .pptx, .jar, .apk, .epub and other files built on ZIP. It is read where it is, not uploaded.

A ZIP extractor that opens archives in your browser

A ZIP keeps a list of its files at its end: each name, size, checksum and where its data begins. This page reads that list first, so it shows what an archive holds at once, however large it is, and then reads only the files you ask for, each in pieces, straight from the file on your disk. Every file it extracts is checked against the CRC-32 checksum the archive keeps, so a damaged download is caught rather than saved. The archive never leaves your computer, and once the page has loaded it works with no network at all.

It was checked against archives made on Windows 10 by 7-Zip, File Explorer, PowerShell, .NET, Python, Java's jar, Windows' own tar and git, and against Office files, a Python wheel and a NuGet package: every file came out byte for byte as 7-Zip extracts it. Names written in UTF-8 show as written, in any script. A ZIP made by Windows' own "Compressed folder" keeps names in the code page of the Windows that made it, such as Shift_JIS on a Japanese one, so the page reads those in the code page of your browser's language, and lets you choose another.

The page will not write outside the folder you choose: a name such as ../../evil.exe, the trick known as zip slip, is refused. Nor will it open an archive whose entries share the same compressed bytes, the trick of a zip bomb, or read a file past the size its archive gives. Entries encrypted with a password, and those compressed with methods other than Deflate, such as Deflate64, LZMA or BZIP2, are listed with the reason the page cannot extract them; 7-Zip opens them. To make a ZIP instead, use the ZIP file compressor.

What this tool is used for

  • Opening a ZIP on a work or school computer where installing 7-Zip or WinRAR is not allowed.
  • Looking at what a ZIP from an email or a download holds before extracting any of it.
  • Taking one file out of a large archive, such as a single photo from a backup, without extracting the rest.
  • Reading a ZIP made by Windows in Japanese, Chinese, Russian or another language, whose names come out garbled elsewhere, by choosing its code page.
  • Looking inside a .docx, .xlsx, .jar, .apk or .epub, which are ZIPs, to see the parts they are made of.
  • Finding out whether a downloaded archive is whole: every file is checked against its CRC-32 as it is extracted, so a damaged one is named rather than saved.

How it works in practice

A worked example

A colleague in Tokyo zips a folder with Windows' Send to > Compressed (zipped) folder and emails it. Opened elsewhere, its Japanese file names come out as strings of accented letters and box symbols.

Input
Names without the UTF-8 flag: DOS, US (code page 437)
ô·û{îΩé╠âtâ@âCâï.txt          41 B
âtâHâïâ_/ÆåÉg.txt              36 B
ascii.txt                      22 B
Output
Names without the UTF-8 flag: Japanese (Shift_JIS)
日本語のファイル.txt           41 B
フォルダ/中身.txt              36 B
ascii.txt                      22 B

A ZIP can flag a name as UTF-8, and then every system reads it the same way. Windows' own Compressed folder does not: it writes each name in the DOS code page of the Windows that made it and sets no flag, so nothing in the archive says which code page that was. Measured on a US Windows 10, it wrote café.txt with é as byte 0x82, code page 437; a Japanese Windows uses Shift_JIS. Read in code page 437, Shift_JIS bytes come out as the accented letters and symbols of the first listing. The bytes themselves are intact, so choosing Shift_JIS under Names without the UTF-8 flag gives the names back, and the files extract under them. In a browser set to Japanese, the page reads such names in Shift_JIS from the start.

The edge case that catches people

Windows' own Compressed folder cannot write a name outside its code page at all: on a US Windows 10, given a file named 日本語.txt, it wrote nothing. Windows' tar.exe writes such characters as question marks, which no code page brings back. And 7-Zip reads names in an archive marked as made on Unix as UTF-8: it showed tar's café.txt as caf.txt, where the page, finding the bytes were not UTF-8, read them in code page 437 as café.txt.

When not to use this tool

For RAR and 7z archives, split archives, and ZIPs encrypted with a password or compressed with Deflate64, LZMA, BZIP2 or PPMd, use 7-Zip: browsers decompress only Deflate, and the page does not decrypt. To make a ZIP rather than open one, use the ZIP file compressor.

Frequently Asked Questions

Is my ZIP uploaded anywhere?
No. The page reads the ZIP where it is on your disk, a piece at a time, with its own JavaScript in your browser tab. Nothing is sent to a server, and once the page has loaded it works with no network at all. A file you save is handed to your browser like any download.
How do I extract every file at once?
In Chrome or Edge, click Extract to a folder and choose one: the page writes every file into it with the archive's folders, inside a new folder named after the ZIP unless you untick that. Firefox and Safari cannot write into a folder, so there each file is a download of its own, all at once for up to 100.
Why do some file names look garbled?
A ZIP made by Windows' own Compressed folder keeps names in the code page of the Windows that made it, with nothing to say which: measured on Windows 10, it wrote café.txt in code page 437. The page reads such names in the code page of your browser's language; if they look wrong, choose another.
Can it open a password-protected ZIP?
No. It lists the files of an encrypted ZIP, ZipCrypto or AES, and says it cannot decrypt them; it never asks for a password. Entries compressed with Deflate64, BZIP2, LZMA or PPMd, which 7-Zip can write, are listed the same way, as browsers decompress only Deflate. 7-Zip opens both.
Is it safe to open a ZIP from someone else?
The page refuses a name such as ../../evil.exe, which would write outside the folder you choose (zip slip); refuses an archive whose entries share the same compressed bytes, a zip bomb's trick; and stops reading any file past the size its archive gives. It never runs a file from the archive.
How large a ZIP can it open?
It reads the list at the archive's end first: one of 70,000 files was listed in Chrome in under a second. Then it reads each file in pieces and writes it straight into the folder, so no file has to fit in memory: a 4.5 GB file was extracted in 54 seconds, with the page's JavaScript memory under 20 MB.
Does it open .docx, .xlsx, .jar or .apk files?
Yes: those are ZIPs, as are .pptx, .epub, .whl and .nupkg. Measured on Word, Excel and PowerPoint templates, a Python wheel, a NuGet package and a Java archive, every file came out as 7-Zip extracts it. A ZIP inside a ZIP has an Open button, which shows what it holds the same way.
Why do extracted files not keep their dates?
A browser gives a page no way to set a file's date when it writes one, so each file carries the time it was written. The list shows each file's date as the archive keeps it.

Common errors and gotchas

  • Taking a .rar or .7z file for a ZIP. Those are other formats, which the page names when given one; 7-Zip opens them.
  • Opening one part of a split archive, a .zip with .z01 parts or a .zip.001. Put every part in one folder and open the first with 7-Zip, which joins them.
  • Waiting for a password prompt. The page does not decrypt an encrypted ZIP; it lists its files and says so.
  • Extracting into the Downloads folder itself. Leave the new folder ticked, so an archive of hundreds of files lands in a folder of its own.
  • Taking garbled names for a broken archive. Names written by Windows' own zipping are in its language's code page; choose it under Names without the UTF-8 flag.

Related Encoders & Decoders tools

Private & free — this tool runs entirely in your browser.

IndieKit — Ship your Next.js startup in days.affiliate