Initializing... drag & drop files here
Supports: TAR.GZ
.tar.gz or .tgz archive, or click "+ Add Files" to browse. Batch uploads work — queue several archives and they all decompress in one pass..tar back). If you upload several archives, switch to Individual Archives to get a separate .tar per input, or keep Single Archive to merge their contents into one master tarball..gz suffix stripped (backup.tar.gz → backup.tar). No re-archiving happens — file headers, permissions, ownership, and timestamps inside the tar are preserved bit-for-bit..tar archive ready for tools that don't support gzip streams.A .tar.gz is two formats stacked: tar bundles files into one stream, then gzip (DEFLATE, RFC 1952) compresses that stream. Converting to plain .tar removes the gzip layer while keeping the archive structure intact. Reasons people do this:
.tar.gz your tool must decompress everything before it. A plain .tar lets tools like tar --list, bsdtar -tvf, or libraries that mmap the archive seek directly to a file's header without scanning prior blocks.xz -9 or zstd --ultra -22 for 20-40% smaller archives on source code and logs. You can't do that in one step without an intermediate decompression.docker load accepts both, but custom OCI tooling sometimes does not).gunzip reports a CRC mismatch but partial data decompressed, saving the recovered bytes as a plain .tar lets tar --ignore-zeros salvage the readable members.tar -tvf archive.tar runs instantly on an uncompressed archive; on a 2 GB .tar.gz it can take 30+ seconds because every byte must pass through the gzip inflater.| Property | TAR.GZ (.tar.gz / .tgz) |
TAR (.tar) |
|---|---|---|
| Compression | DEFLATE via gzip (RFC 1952) | None — raw concatenated file records |
| Typical size on source code | ~25-35% of uncompressed | 100% (baseline, often slightly larger due to 512-byte block padding) |
| Random access | No — must decompress sequentially | Yes — tool can seek to any file header |
| List contents speed (2 GB archive) | 15-45 seconds (full inflate) | Under 1 second (header walk only) |
| Native on Windows 11 | Yes (22H2+, via libarchive) | Yes (tar.exe ships in build 17063+) |
| Native on macOS / Linux | Yes (tar -xzf) |
Yes (tar -xf) |
| Stores POSIX permissions | Yes | Yes |
| Stores symlinks / hardlinks | Yes | Yes |
| Resilient to bit-flip | Low — one bad byte kills the rest of the stream | High — corruption is local to that file's blocks |
| You want to... | After this conversion you can... |
|---|---|
| Re-compress with xz | Run xz -9 archive.tar to get a smaller .tar.xz (try TAR to TAR.XZ) |
| Re-compress with bzip2 | Run bzip2 archive.tar or use TAR to TAR.BZ2 |
| Switch to ZIP for Windows users | Use TAR to ZIP — adds CRC-32 per file and random-access central directory |
| Extract individual files fast | Open the .tar with bsdtar, 7-Zip, or The Unarchiver — seek is instant |
| Reverse the conversion | Use TAR to TAR.GZ to re-apply gzip compression |
Yes — that's expected. Gzip removes redundancy in the tar stream; converting back to plain tar restores the uncompressed size. A 50 MB .tar.gz of source code often expands to 200-300 MB as .tar. If size matters, re-compress with .tar.xz or .tar.zst after conversion.
.tgz the same as .tar.gz?Functionally identical. .tgz is a shortened form for filesystems that historically disliked double extensions (old DOS/Windows 8.3). The bytes inside are the exact same gzip-wrapped tar stream, so our converter accepts both and treats them the same way.
No. Only the outer gzip wrapper is removed. File names, directory structure, POSIX permissions (mode bits), UID/GID ownership, modification timestamps, symlinks, and hardlinks inside the tar are preserved byte-for-byte. Run sha256sum on individual files extracted before and after — they match.
gunzip locally?You can — gunzip backup.tar.gz produces backup.tar in one step. This online tool is for users who don't have a terminal handy (Windows users on older builds, Chromebooks without Linux containers, locked-down corporate machines, or mobile devices). On Windows 11 22H2 or later you also have tar.exe built in; on macOS and Linux, tar and gunzip are preinstalled.
Standard tar.gz does not support passwords — gzip has no encryption layer. If your file was encrypted, it was either wrapped in something else (GPG, OpenSSL, age) or repackaged as .7z / .zip with AES. Decrypt with the original tool first, then upload the resulting .tar.gz here.
Gzip's solid stream means a CRC error usually halts decompression at the bad byte. Our converter returns whatever decompressed successfully before the error. To squeeze more data out, download the partial .tar and run tar --ignore-zeros -xvf partial.tar locally — tar will skip past damaged regions and recover the surviving file members.
Free users can convert archives up to several GB; the exact ceiling depends on browser memory for the initial upload chunk. For multi-GB scientific datasets or full Linux distro tarballs, the command line (gunzip or tar -xzf) will always be faster than any browser-based tool because it streams without uploading.
If you store backups as .tar.gz and only need to restore one file, you currently decompress the entire archive. Converting to .tar first lets indexing tools (like tar --index-file= or mkindex.tar) build an offset table so future single-file extractions take milliseconds instead of seconds-to-minutes.
tar -xzf?No — tar -xzf tells tar to expect gzip-compressed input. After conversion, use tar -xf (no z flag) or tar -xvf for verbose listing. Most modern tar implementations auto-detect compression, so tar -xf archive.tar works whether the input is compressed or not, but explicitly dropping the z is cleaner.