Initializing... drag & drop files here
Supports: TAR.XZ
.tar.xz (or .txz) archive, or click "+ Add Files" to browse. Batch upload is supported — drop several archives in at once..tgz. Switch to Individual Archives to get a separate .tgz per input — useful when you want each release artifact to stay isolated.myproject-1.4.tgz). The repack preserves the inner TAR's file paths, permissions, and timestamps — only the outer compression layer changes..tgz (a.k.a. .tar.gz).TAR.XZ wraps a tar archive in XZ Utils LZMA2 compression (specification published January 14, 2009), while TGZ wraps the same tar payload in gzip DEFLATE compression (released October 31, 1992 by Jean-Loup Gailly and Mark Adler). The TAR contents — file paths, permissions, ownership, timestamps — are identical; only the outer compression layer differs. Going XZ → gzip trades a smaller file for a faster, more universally supported one.
gzip but not xz. A .tgz extracts with tar -xzf on any POSIX system going back decades; tar -xJf needs xz-utils installed..tgz is the lowest-friction format to drop into a build pipeline..tgz for its packages before switching to .txz in version 13.0 (2009); converting back to .tgz keeps the package readable by older Slackware utilities and third-party tools that never adopted XZ.Content-Encoding: gzip), so a .tgz can be streamed and decompressed inline. XZ has no equivalent transport encoding.xz-utils, some teams locked their build pipelines to gzip-only. Repacking inbound .tar.xz archives as .tgz is one mitigation step.| Property | TAR.XZ | TGZ (TAR.GZ) |
|---|---|---|
| Outer compression | LZMA2 (XZ Utils, 2009) | DEFLATE (gzip, 1992) |
| Typical ratio (Linux kernel src) | ~14.4% of original | ~21.3% of original |
| Compress time at level 6 | ~383 s | ~31 s |
| Decompress time | ~11 s | ~6 s |
| Memory at decompress | Up to ~65 MB (default preset) | ~256 KB |
| Integrity check | CRC-32, CRC-64, or SHA-256 | CRC-32 |
tar flag to extract |
tar -xJf (needs xz-utils) |
tar -xzf (built-in) |
| Streaming HTTP support | No standard Content-Encoding | Content-Encoding: gzip |
| Best when | Distribution size matters | Decompression speed and compatibility matter |
Benchmarks above from RootUsers gzip-vs-xz comparison on a 580 MB source corpus at each tool's default compression level.
| Scenario | Keep as TAR.XZ | Convert to TGZ |
|---|---|---|
| One-time download by many users | — | Yes (lower CPU at extract) |
| Long-term cold storage | Yes (smaller on disk) | — |
| Distributing to older / embedded Linux | — | Yes (gzip is ubiquitous) |
| Building a Docker image | — | Yes (matches default layer compression) |
| Linux distro ISO / package | Yes (size wins on slow mirrors) | — |
| Sharing with non-technical Windows users | — | Yes (more tools handle .tar.gz than .tar.xz) |
| CI artifact for fast restore | — | Yes |
No. The inner TAR stream is identical — file paths, Unix permissions (mode bits), owner/group IDs, and modification timestamps all carry over byte-for-byte. Only the outer compression layer is swapped from LZMA2 to gzip's DEFLATE, so tar -tzf output.tgz will list the exact same entries as tar -tJf input.tar.xz.
For typical source-code or document archives, expect the .tgz to be roughly 40-60% larger than the .tar.xz. On the canonical 580 MB Linux kernel benchmark, gzip compressed to 21.3% of original vs XZ's 14.4% — so the gzip output is about 1.5x the XZ size. Heavily binary or already-compressed payloads (JPEGs, MP4s) will show a smaller gap because neither algorithm compresses them well.
.tgz the same as .tar.gz?Yes, exactly the same — .tgz is just the short-form extension. Both filenames decode to a gzip-compressed TAR archive with identical bytes. The longer .tar.gz is more common on Linux servers; the four-letter .tgz was originally used because some older filesystems (notably FAT on DOS/Windows) only supported one extension at a time.
Yes. Run the reverse conversion at tgz to tar.xz. The recompression is lossless in the sense that the inner TAR is preserved, but you'll lose XZ's integrity checksums (CRC-64 / SHA-256) from the original and pick up gzip's CRC-32 instead. If those checksums matter, keep a copy of the original.
xz-utils installed to do this conversion?Not on xconvert — the LZMA2 decompression and gzip recompression both run server-side, so your machine only needs a browser. If you wanted to do it locally instead, the one-liner is xz -dc input.tar.xz | gzip -9 > output.tgz, which requires both xz-utils and gzip on your PATH.
Decompressing the LZMA2 layer is the slow step — XZ decompression is roughly 80% slower than gzip decompression at default settings, and it has to fully decode the stream before gzip can re-encode the TAR. For a 100 MB .tar.xz you should expect 5-15 seconds depending on server load; the gzip recompress itself is near-instant.
The default repack targets level 6 — gzip's standard, which is what most distros and CI tools produce by default. This gives a good balance: typically within 1-2% of gzip's maximum (level 9) compression ratio, at roughly half the CPU cost. The output is byte-compatible with any gunzip, tar -xzf, or Content-Encoding: gzip consumer.
.tgz?Yes. 7-Zip, PeaZip, and WinRAR all handle .tgz and .tar.gz natively, and Windows 10/11 ships with tar in PowerShell (use tar -xzf file.tgz). macOS Finder double-click also unpacks .tgz via the built-in Archive Utility. Compatibility is one of the main reasons to convert from .tar.xz in the first place.
Yes — see TAR to TGZ for an uncompressed TAR, TAR.GZ to TGZ (rename / repack), 7Z to TGZ, or RAR to TGZ. All variants preserve directory structure and produce a standard gzip-compressed TAR on the output side.