Open a TAR or TAR.GZ file
Unpack the archives that come out of Linux and macOS without opening a terminal. TAR, TAR.GZ, TGZ and plain GZ are all supported, and nothing is stored.
- Never stored
- No queue, no waiting
- No signup, no watermark
How it works
Add your archive
TAR, TAR.GZ, TGZ or a single gzipped file. Compressed ones are decompressed first, automatically.
Look inside
Every file is listed with its size and folder. Links are shown as links rather than followed.
Take what you want
Download one file, or all of them together as a ZIP that opens anywhere.
Two formats stacked, and what a download cannot restore
A .tar.gz is two formats in a stack, and nearly everything about TAR follows from that. TAR itself compresses nothing: it concatenates files into one stream, each preceded by a 512-byte header holding the name, size, permissions, owner and timestamp. The compression is a separate layer wrapped around the whole thing — gzip, bzip2 or xz — which is why the extensions stack up and why the compression ratio is better than a ZIP's: ZIP compresses each file separately, so it cannot exploit repetition between files, while a compressed TAR squeezes the entire stream as one.
The same stacking explains TAR's one real weakness. Because the compression wraps everything, there is no catalogue to read: listing the contents of a .tar.gz means decompressing it from the beginning, and getting the last file out means decompressing all of it. A ZIP lists instantly however large it is. This is a design trade from an era of magnetic tape, where reading sequentially was the only option there was — hence the name.
Unix permissions and symbolic links are listed and cannot be restored, and this is the limit that matters most for what people use TAR for. Downloads arrive as ordinary files with ordinary permissions, so an executable script comes out non-executable and a symbolic link comes out as the small file it physically is, containing the path it pointed at. For unpacking a source release you intend to build, that is a real obstacle rather than a detail — the metadata is read and shown here, but a download cannot carry it.
Ownership is worth a note too. TAR records the user and group that owned each file, which is exactly why extracting a TAR as root on a server is a thing people are warned about. None of that applies to a download, which is the one respect in which this is safer than the command line.
When you want something else
For anything you intend to use rather than merely look at, tar itself is the answer and it is on every Mac and Linux machine already: tar -xzf archive.tar.gz restores permissions, symbolic links, ownership and timestamps, extracts straight to disk with no memory ceiling, and handles archives of any size. Windows has had it built in since 2018, so tar -xf works in a plain command prompt.
The specific cases this cannot help with are the ones TAR exists for: unpacking a source tarball to compile, restoring a backup, or moving a directory tree between machines with its structure intact. All three depend on exactly the metadata a download discards. This page is for seeing what is in an archive, or pulling out a document or an image on a machine where you cannot install anything.
Frequently asked questions
What is the difference between .tar and .tar.gz?
A .tar is not compressed at all — it just glues files together into one, keeping their names and folders. A .tar.gz is that same bundle with gzip wrapped around the whole thing. Gluing first and compressing second is why .tar.gz is often SMALLER than a ZIP of the same files: gzip sees every file as one continuous stream and can spot repetition across the boundaries between them, where a ZIP compresses each file separately and starts afresh every time.
Is my archive stored anywhere?
No. Nothing is stored and nothing is kept, and it keeps working with the Wi-Fi switched off once the page has loaded. That matters more for archives than for most files — a zip of your tax returns or a backup of a folder is exactly the kind of thing that should not be handed to a stranger's server.
Do long file names work?
Yes. The original tar format could not store a name longer than 100 characters, and two incompatible extensions grew to fix that — GNU's and POSIX pax. Both are supported, so an archive from any real system unpacks with its names intact.
What happens to symbolic links?
They are listed, with what they point at, but never followed. Following a link inside an archive is one of the classic ways an extractor gets talked into writing somewhere it should not, so it is simply not done.
Is there a size limit?
Available memory, rather than a fixed size cap. Very large archives are opened a file at a time where the format allows it, so a big ZIP lists instantly and only the file you ask for is unpacked.
Is it safe to open a TAR from someone I do not know?
Archives are the one kind of file whose CONTENTS can be hostile, so two old attacks are handled before you see anything. A file inside an archive can be named "../../something" to try to escape the folder it is unpacked into — those names are clamped and you are told which ones were changed. And an archive can be a few kilobytes that unpacks to terabytes, which would take the browser tab down with it; that is measured and refused rather than attempted.
Good to know: TAR, TAR.GZ, TAR.BZ2 and TAR.XZ all open — the last three through libarchive, which is fetched only when a compressed archive is actually opened, so the plain TAR path stays small. Unix permissions and symlinks are listed but cannot be restored: a browser has no access to either.
Put this tool on your website
Free for any blog, class page or help article. Paste one snippet and your visitors can use it right on your page.