Initializing... drag & drop files here
Supports: TAR.BZ2
.tar.bz2 (or .tbz2 / .tb2) archives from your device. Batch upload is supported — recompress a folder of old backups, kernel sources, or CRAN package mirrors in a single pass..tgz per uploaded .tar.bz2 (the default for archive-to-archive jobs), or switch to Single Archive to merge every upload into one combined .tgz. Internal compression level uses gzip's default -6 setting..tgz extension is just the legacy 8.3-filename short form. Either name extracts with tar -xzf on Linux/macOS or 7-Zip / WinRAR on Windows without any flag change..tgz individually or as a bundled ZIP. No sign-up, no watermark, no file-count cap..tar.bz2 (bzip2-compressed tarball) was a common Linux distribution format from the late 1990s through the mid-2010s because bzip2 squeezes about 15-20% more out of source code than gzip. The trade-off is speed: gzip's DEFLATE algorithm decompresses roughly 3-6x faster than bzip2's Burrows-Wheeler transform on the same hardware, and that gap matters every time a CI pipeline or a user has to unpack the file. Recompressing to .tgz keeps the same tar archive structure (permissions, symlinks, ownership, mtimes) but swaps in gzip for faster extraction at minor size cost.
.tgz typically lands in 4-6 seconds, which compounds across hundreds of CI jobs per day.gzip shipped with virtually every Unix since the early 1990s, and Windows tools like 7-Zip, PeaZip, and the built-in tar.exe in Windows 10+ all read .tgz natively. bzip2 support is universal on Linux/macOS but more variable on locked-down corporate Windows installs.curl... | tar -xz for one-shot download-and-extract on bandwidth-constrained boxes. bzip2 works the same way in principle but is significantly more CPU-bound, which slows down small VPS instances..tgz drops the CPU cost with negligible size penalty..tar.gz (Debian source packages, Node.js releases, Python sdist) or .tar.xz (the GNU project's archives since 2014, kernel.org). .tar.bz2 is increasingly legacy; converting old mirrors to .tgz aligns with what current tooling expects..tgz they can open with the built-in tar command beats sending .tar.bz2 and asking them to install extra tooling.Going the other direction or to a different format? See TGZ to TAR.BZ2 for the reverse, TAR.BZ2 to TAR.GZ (the long-form equivalent of .tgz), or TAR.BZ2 to ZIP if your recipient is on Windows without tar. To just unpack the contents, use Extract TAR.BZ2 or Extract TGZ.
| Property | TAR.BZ2 (.tar.bz2 /.tbz2 /.tb2) | TGZ (.tgz /.tar.gz) |
|---|---|---|
| Archive layer | POSIX tar (ustar / pax) |
POSIX tar (ustar / pax) |
| Compression algorithm | bzip2 — Burrows-Wheeler transform + Huffman coding | gzip — DEFLATE (LZ77 + Huffman) |
| Author / year | Julian Seward, 1996 | Jean-loup Gailly & Mark Adler, 1992 |
| Typical compression ratio (source code) | ~15-20% smaller than gzip | Baseline |
| Compression speed (rootusers Linux-kernel test) | ~83 sec at bzip2 -9 | ~31 sec at gzip -6 |
| Decompression speed (same test) | ~23 sec | ~6 sec |
| Block size | 100-900 KB blocks | 32 KB sliding window |
| Native CLI on macOS/Linux | tar -xjf (or -xf auto-detect) |
tar -xzf (or -xf auto-detect) |
Windows 10/11 native tar.exe |
Yes (libarchive-backed) | Yes (libarchive-backed) |
| Corruption recovery | Block-level — bzip2recover salvages intact blocks |
Stream-level — corruption past offset typically forfeits rest |
| Common modern uses | Legacy Linux source mirrors, FreeBSD ports tarballs | Node.js / Python / Ruby source releases, Docker layers (older) |
| Tool | Algorithm | Default level | Strength | Weakness |
|---|---|---|---|---|
gzip (.tgz / .tar.gz) |
DEFLATE | -6 |
Fastest of the three; universal support | Largest output of the three |
bzip2 (.tar.bz2 / .tbz2) |
Burrows-Wheeler + Huffman | -9 |
Smaller than gzip; recoverable blocks | 3-6x slower than gzip on decompress |
xz / lzma (.tar.xz) |
LZMA2 | -6 |
Smallest output (typically 25-30% better than gzip on source) | Slowest compress; high memory at high levels |
zstd (.tar.zst) |
Zstandard | -3 |
Near-gzip size at far better speed; multi-threaded | Newer (2016); slightly less ubiquitous on legacy Unix |
If you're recompressing for size rather than speed, consider .tar.xz instead — see TAR.BZ2 to TAR.XZ. If you need the broadest Windows compatibility without any tar tooling, use TAR.BZ2 to ZIP.
Yes — byte-for-byte identical content, just two extensions for the same gzip-compressed tarball. The .tgz short form dates back to the 8.3 filename limit on MS-DOS and early Windows, where .tar.gz ran out of room. Both extract with the same command (tar -xzf file.tgz or tar -xzf file.tar.gz) and both are valid in modern tooling. If you need the long-form name instead, use TAR.BZ2 to TAR.GZ — it produces the same bytes with the alternate suffix.
No file contents are lost. The conversion fully decompresses the bzip2 stream into the original tar archive, then recompresses the same tar bytes with gzip. File names, permissions (chmod bits), ownership (UID/GID), symbolic links, hard links, modification times, and extended attributes inside the tar layer all survive the round-trip. The only thing that changes is the outer compression wrapper.
Typically 10-20% larger for plain text or source code, sometimes negligible for content that's already compressed (JPEGs, MP4s, PDFs). In the rootusers Linux-kernel benchmark, the same source tree compressed to 96 MB with bzip2 -9 versus 123 MB with gzip -6 — about a 28% size bump going from bzip2 to gzip on a worst-case text-heavy payload. Real-world deltas on mixed archives are smaller.
Decompression speed. Bzip2's Burrows-Wheeler transform is roughly 3-6x slower to unpack than gzip's DEFLATE on the same hardware. For a 1 GB archive that gets pulled by every CI build, every container image bake, or every server provision, the extraction time difference dominates the small bandwidth savings. Gzip is also more universally supported on locked-down corporate Windows boxes where bzip2 isn't part of the default toolchain.
Yes. The conversion operates on the outer compression layer only — the inner tar archive is byte-identical between the input .tar.bz2 and the output .tgz. POSIX permissions, ownership, symlinks, hard links, sparse-file metadata, and pax extended headers all pass through unchanged. If you extract the .tgz on a system that supports those attributes, you'll get the same tree the original .tar.bz2 would have produced.
All three are the same bzip2-compressed tarball — .tbz2 is the 4-character short form, .tb2 is a 3-character variant used on legacy 8.3 filesystems, and .tar.bz2 is the verbose form. This converter accepts any of them. Output is .tgz by default; for the long-form .tar.gz name, use TAR.BZ2 to TAR.GZ.
No — that produces a corrupt archive. The extension labels the compression wrapper; the bytes inside .tar.bz2 are bzip2-compressed, and tar -xzf (gzip mode) will refuse them with a "not in gzip format" error. You have to actually decompress with bzip2 and recompress with gzip, which is what this tool does in one step.
Windows 10 (October 2018 update / 1809) and Windows 11 ship tar.exe built in. Open Command Prompt or PowerShell and run tar -xzf yourfile.tgz. For a graphical option, 7-Zip and PeaZip both read .tgz natively. If you'd rather not deal with tar at all, convert to ZIP instead with TAR.BZ2 to ZIP — File Explorer's built-in "Extract All" handles ZIPs directly.
files are processed on our servers and deleted automatically after a few hours and deleted after the session ends. No account is required and there are no watermarks, file-count limits, or hidden Pro tiers. If you only need to peek inside the original .tar.bz2 without recompressing, use Extract TAR.BZ2 to pull out the contents directly.