Zip Tools: Create and Extract Archives in Your Browser

By tobeannounced · Carteret, NJ

Most people meet archive formats one of two ways: a folder that has to go out as a single file, or a download that arrives as something the system refuses to open. This zip tool handles both directions in the browser. It writes archives in 23 formats and reads 57 of them, and it does the work on your own machine instead of a server.

What a zip tool actually does

Two different jobs hide behind the word zip, and mixing them up explains most of the confusion around .zip, .tar, .gz, .tar.gz, .7z and .rar.

Bundling means packing many files, with their names and folder structure, into one container. Compression means taking one stream of data and making it smaller. Some formats do only one of the two. TAR bundles and does not compress at all, which is why a .tar is often as large as the files inside it. GZIP compresses exactly one stream and bundles nothing, which is why a .gz holds a single decompressed file. ZIP, 7z and RAR do both at the same time.

That split is also why Unix systems invented .tar.gz. TAR collects the tree first, then gzip squeezes the result. When you drop a .tar.gz here, the tool reports both parts: tar + gzip. It is telling you which decoder it ran, not guessing from the filename.

What this tool opens: 57 formats, one drop zone

The extract tab takes one archive at a time, of any size, and you never pick the format. The tool reads the first bytes of the file and decides for itself.

There is a signature for ZIP, TAR, gzip, bzip2, xz, zstd, lzma, lz4, snappy, 7z, rar, cab, iso, cpio, ar and deb packages, rpm, warc, lzip, mtree, lha, xar, compress .Z, wim, vhd, vhdx, vdi, vmdk, qcow, dmg, apfs, ext, ntfs, fat, hfs, squashfs, cramfs, Intel HEX, gpt, mbr, uefi firmware, chm help files, compound documents such as msi, doc and xls, nsis installers, and Electron asar bundles. Twenty-four of those are handled by codecs built into the page. The remaining thirty-three go to two bundled engines, libarchive and 7-Zip, plus a small asar reader.

A handful of formats carry no signature at all. A bare DEFLATE stream, an LZNT1 block, an LZString payload, a Snappy block, base64 text and uuencoded text all look like ordinary bytes. Those six are decided by the file name you drop, and a successful extraction is the proof.

When the tool reads an archive it prints what it found (zip, tar + gzip, dmg, compound file (msi/doc/xls)) and lists every entry with its size. You can pull out a single file without writing the whole archive to disk first.

What this tool creates: 23 formats

The create tab is the other half. Add files, choose a format, download the result. These are all standard formats, written by libraries rather than by a hand-rolled encoder.

TAR.LZMA, TAR.LZ4, TAR.BZ2 and TAR.BR cover the remaining tar and compressor pairings, and LZMA, BZIP2, Snappy, LZString and raw DEFLATE and ZLIB are available as standalone streams. LZNT1 is the exception: the tool can unpack it but has no encoder for it, so it does not appear in the create list.

Creating an archive, step by step

1. Open the create tab and drop your files into the box. Files line up in a table with their sizes, and any of them can be removed before you continue.
2. Pick a format from the dropdown. The button label changes to match, so you can see what you are about to get.
3. Press the button. A progress bar runs while the archive is built in memory.
4. The finished file downloads with the right extension.

One thing catches people out. Stream formats wrap exactly one file: GZ, XZ, LZ4, ZSTD, BROTLI, BZIP2 and the rest, plus the base64 and uuencode wrappers. Add two files and the tool answers single file only, then stops. That is correct behaviour, not a bug: a .gz has room for one stream. Formats that hold many files are ZIP, TAR, the tar plus compressor combinations, 7Z and WIM.

For a folder of documents going to a client, ZIP is the answer. For a build artifact on a Linux box, TAR.GZ or TAR.ZST. For the smallest possible file when you know the other side has 7-Zip, 7Z.

Opening an archive, step by step

1. Switch to the extract tab and drop one archive.
2. Watch the progress bar. The format label appears when detection finishes.
3. Read the entries table: name and size for every file inside.
4. Download what you need, one entry at a time.

Detection works from content, so a .zip that is really a .7z still opens, and a file with no extension at all still works. Password-protected archives are the case it cannot handle, because the contents are unreadable without the key. There is no password field and no attempt to guess, so an encrypted archive is reported as unsupported rather than half-extracted.

Which format should you pick

The trade is always the same: compatibility against size.

How much smaller things actually get depends on content, not on the format alone. Plain text, CSV files, logs and source code typically shrink by 70 to 90 percent under DEFLATE, and slightly more under LZMA. Files that are already compressed barely move: JPEG, PNG, MP3, MP4, PDF and modern Office documents sit at a few percent, because the repetition has already been removed. Office files are ZIP containers internally, so zipping them again achieves almost nothing.

Worth knowing before you pick 7Z for a letter you are emailing: Windows opens ZIP out of the box, macOS opens ZIP, .gz and .tar.gz, and nothing opens .7z, .rar or .xz without extra software. ZIP is the only format you can assume a stranger can open.

Why nothing gets uploaded

Every step happens in the page. Files are read with the browser's file API, decompressed by codecs compiled to WebAssembly and JavaScript, and handed back as a download. There is no upload, no queue, no account, and no copy of your file on a server.

That matters more for archives than for most tools. An archive is usually a bundle of the things you would least want to hand over: contracts, invoices, source code, database dumps, medical scans. Client-side processing means there is nothing to delete afterwards, because there was never anything to store. It also means the tool works on a metered connection without paying twice for your own data.

The practical ceiling is memory, not policy. A 125 MB disk image fails in the tab, and comfortable work sits in the tens of megabytes. Splitting a heavy job into two smaller archives is usually easier than fighting that limit.

What this zip tool cannot do

Stated plainly, so you do not spend time on a dead end:

• Password-protected archives are not supported in either direction.
• Multi-part archives (.zip.001, .part1.rar, .r01) are not reassembled.
• One archive is extracted per drop.
• ARJ and shar are recognised and refused honestly, instead of failing halfway with a generic error.
• exFAT disk images are not opened: the engine cannot read valid exFAT images, so the format is not offered. FAT reads fine.
• Bare LZIP and raw .lz streams refuse, while .tar.lz works.

Everything on that list has a reason, and none of it is an accident of the build. The formats offered here have been checked on files made by other programs, not merely on this page's own output.

tobeannounced

Carteret, NJ · SO reputation 1999 · Badges: 1🥇20🥈30🥉 · SO member since 2010