Initializing... drag & drop files here
Supports: TAR.XZ
.tar.xz (also seen as .txz) archive. Batch upload is supported — drop multiple archives to decompress them in one pass..tar, which strips the outer XZ/LZMA2 layer while preserving the original tar stream byte-for-byte — filenames, permissions, ownership, symlinks, and timestamps are all retained..tar.xz, choose Single Archive to repackage them into one tar, or Individual Archives (default for archive-to-archive jobs) to emit one .tar per input..tar is ready to extract with tar -xf, 7-Zip, WinRAR, or any POSIX-compliant archiver — no XZ Utils required on the destination machine.A .tar.xz is just a tar archive wrapped in an XZ (LZMA2) compression layer. Converting to .tar removes the compression but keeps the archive structure intact, which is useful whenever you need to inspect, repack, or hand the data off to a tool that can't handle the XZ stream itself.
.tar sidesteps that and lets tar, pax, or BSD bsdtar read the file directly..tar first, then pass through gzip, zstd, or brotli. xconvert can do this in two clicks via TAR.XZ to TAR.GZ..tar is byte-addressable, so backup tools and dedup engines (restic, BorgBackup, ZFS send) can stream and chunk it efficiently..tar.xz and .tar.gz, but rebuilders sometimes need the inner .tar to apply patches or generate reproducible source RPMs..tar (via xz --decompress --keep or tar with --ignore-zeros) lets you salvage entries before the corruption point. A clean .tar is far easier to triage than a half-decoded LZMA2 stream..tar/.tar.gz support in Windows 11 24H2) still can't open .tar.xz. Hand them the .tar instead, or convert to TAR.XZ to ZIP for the friendliest cross-platform experience.| Property | TAR.XZ (.tar.xz, .txz) |
TAR (.tar) |
|---|---|---|
| What it is | A tar archive compressed with the XZ (LZMA2) stream | Uncompressed concatenation of files with headers |
| Compression algorithm | LZMA2 (XZ format released January 2009) | None |
| Typical size for Linux 2.6.32 source | ~53 MB (xz) | ~360 MB uncompressed |
| Random access | No — stream must be decoded sequentially | Yes — headers are byte-addressable |
| Integrity check | CRC-32, CRC-64, or SHA-256 (per XZ spec) | None built into the format itself |
| First standardized | XZ file format spec 1.0.0, Jan 2009 | Unix V7 (1979); POSIX.1-1988 USTAR; POSIX.1-2001 pax |
| Decompression memory | ~8 MiB for default -6 preset (dictionary size) |
Negligible |
| Native Windows 11 support | No (24H2 still excludes .tar.xz) |
Yes (since 24H2, Explorer reads .tar) |
GNU tar flag |
-J / --xz |
(default — no flag) |
| Best for | Distribution, long-term storage of source code | Intermediate processing, streaming, repackaging |
.tar.xz files in the wild are usually produced with one of these xz preset levels. Knowing which was used tells you what to expect when decompressing.
| Preset | Dictionary size | Decompressor RAM | Notes |
|---|---|---|---|
xz -0 |
256 KiB | ~1 MiB | Fastest; rare in distribution tarballs |
xz -3 |
4 MiB | ~5 MiB | Common for CI artifacts where speed matters |
xz -6 (default) |
8 MiB | ~9 MiB | Default for most .tar.xz releases on kernel.org, GNU FTP |
xz -9 |
64 MiB | ~65 MiB | Maximum ratio; used for kernel and toolchain tarballs |
xz -9e (extreme) |
64 MiB | ~65 MiB | Extra CPU on encode; same decode memory as -9 |
Converting to .tar removes this entirely — the resulting file is the same regardless of which preset was used.
Yes, in nearly all cases. XZ is a lossless stream codec, and xz --decompress (which is what this conversion runs under the hood) reproduces the exact byte sequence that was fed into the encoder. The only exception is if the original was created with a non-standard tar format flag (e.g. GNU --format=oldgnu vs --format=pax) — in that case the inner tar layout is whatever the original packager chose, and decompression doesn't change it.
XZ with LZMA2 typically achieves a 60–80% size reduction on source code and text-heavy archives. The kernel.org 2.6.32 tarball ran 53 MB as .tar.xz versus around 360 MB uncompressed — a ~6.8× expansion. For binary blobs that are already compressed (JPEGs, MP3s, pre-compressed PDFs) the ratio is much smaller, often under 1.1×.
No. The XZ layer is real binary compression, not just a file extension. tar -xf on a renamed .tar.xz will fail with "This does not look like a tar archive" because the first bytes are the XZ magic (FD 37 7A 58 5A 00), not a tar header (which starts with a filename in ASCII). You have to actually run xz -d (or this converter) to strip the compression.
They're the same format with different naming conventions. .tar.xz is the long form used by GNU projects and most Linux distributions; .txz is the short form preferred on FreeBSD (where it's also the official package format since FreeBSD 10) and on Windows where 8.3 filenames still occasionally matter. xconvert accepts both.
For distribution, yes — XZ compresses roughly 25–30% smaller than gzip on source code at the cost of significantly slower compression (gzip is 10× faster to encode) and modestly slower decompression (2–3× slower than gzip). For backup and intermediate processing, .tar.gz is usually preferred because gzip's speed matters more than the extra savings. See TAR.GZ to TAR if you have the other format.
Yes. All of that metadata lives inside the tar headers, which sit underneath the XZ compression layer. Decompressing to .tar doesn't touch the headers — file modes, uid/gid, mtime, symlinks, hard links, and extended attributes (if stored in pax or GNU format) all come through unchanged.
Windows 11 version 24H2 (released October 2024) added native .tar extraction to File Explorer. Older Windows versions need 7-Zip, WinRAR, or PeaZip — all of which are free and have handled .tar for over a decade. macOS and every Linux/BSD ships tar by default.
Yes, with TAR to TAR.XZ. If you want gzip compression instead (faster, larger), use TAR to TAR.GZ. And if your destination supports newer codecs, TAR.XZ to 7Z gives you a single-file archive with similar ratios and broader Windows support.