Extract huge archives in place. The archive shrinks as your files hatch,
so you only need free space for the output, not for the output and the archive.

A 180 GB game download that unpacks to 200 GB needs 380 GB of disk with WinRAR, 7-Zip or unzip, because the archive and everything extracted from it have to exist at the same time. HatchZip needs a little over 200 GB. It extracts the archive from the end and cuts each finished part off the archive file as it goes, so the space the archive gives back is reused for the next files.

It checks that the extraction will fit before it changes anything, keeps a safety margin of free space the whole way through, and can be stopped or crash at any moment and pick up where it left off. It reads ZIP, 7z, tar, .tar.gz and split archives.

No coding needed. Download for Windows (Windows 10 and 11), or for macOS (Apple Silicon), macOS (Intel) and Linux.

1. Run the installer. Windows may show "Windows protected your PC" because HatchZip is not code-signed yet: choose More info, then Run anyway.
2. Right-click the archive, choose HatchZip, then Extract here. On Windows 11 the menu is under Show more options.
3. Or open HatchZip from the Start menu, open the archive, and choose Extract.

HatchZip shows whether the files will fit before it changes anything. The archive is used up as the files come out, so when it finishes, only the extracted files are left.

The same 200 MB of game data (21 files), packed by the real tools, extracted on a 220 MiB disk with about 50 MB free:

Measured on Fedora 43, Intel Core i7-14700HX, SK Hynix NVMe SSD, warm page cache; the disk-space runs used a 220 MiB tmpfs to stand in for a full laptop disk, the speed runs ext4. Speed comes from decompressing files in parallel, so an archive holding one giant file extracts at the speed of a single core.

The extra space HatchZip needs beyond the output is about the size of the largest single unit it can free at once: one file for ZIP and tar, one solid block for 7z. On filesystems that can punch holes (ext4, XFS, btrfs, tmpfs; NTFS and APFS not yet verified), a large file stored or deflated in a ZIP, or in a tar, frees its archive bytes every 256 MiB while it is extracted: a ZIP holding one 600 MiB stored file extracted on a 950 MiB tmpfs with 367 MB free, where it needed 629 MB before. hatchzip check computes the exact figure for any archive.

Download from the latest release:

The installers also put the hatchzip command (and its short name hz) on your PATH. The Windows setup and the macOS apps are not code-signed yet: on Windows choose More info, then Run anyway; on macOS 15 and later, open the app once, then allow it under System Settings, Privacy & Security.

Linux and macOS:

curl -fsSL https://raw.githubusercontent.com/xevrion/hatchzip/main/install.sh | sh

Windows, in PowerShell:

irm https://raw.githubusercontent.com/xevrion/hatchzip/main/install.ps1 | iex

Both download the build for your machine from the latest release, check it against its published SHA-256 checksum, and install hatchzip and hz for your user only, with the man page and shell completions. No root or administrator rights.

cargo install hatchzip

This builds the command line tool from crates.io. The engine is a library too: hatchzip-core.

With Rust (the version is pinned in rust-toolchain.toml and installed automatically):

git clone https://github.com/xevrion/hatchzip.git
cd hatchzip
cargo build --release

The binaries are target/release/hatchzip and target/release/hatchzip-gui (.exe on Windows). Building the desktop app on Linux needs the fontconfig headers (libfontconfig1-dev or fontconfig-devel).

hatchzip completions bash > ~/.local/share/bash-completion/completions/hatchzip
hatchzip completions zsh > "${fpath[1]}/_hatchzip"
hatchzip completions fish > ~/.config/fish/completions/hatchzip.fish
hatchzip completions powershell >> $PROFILE

hatchzip man > ~/.local/share/man/man1/hatchzip.1    # then: man hatchzip

hatchzip check game.zip

game.zip  180.2 MB zip, 200.0 MB uncompressed in 22 entries
  available                    50.5 MB
  standard extraction needs    200.0 MB
  required                     29.9 MB (+5.0 MB reserved)
  available after              30.6 MB

check reads the archive's index and the disk's free space and changes nothing.

hatchzip extract game.zip                 # into ./game/
hatchzip extract game.zip -o D:\Games\Foo # somewhere else
hatchzip x game.7z.001                    # split archive; any part's name works

HatchZip prints the same space plan, asks for confirmation, then extracts. When the extraction is complete, the archive has been removed and the files are in the output folder. For a split archive, each volume is deleted as soon as everything in it is extracted.

Press Ctrl+C and HatchZip finishes the files it is writing, saves its progress and exits. A second Ctrl+C quits immediately, which is also safe. To continue, run the same command again:

hatchzip extract game.zip

The same applies after a crash, a power cut, or running out of space because something else filled the disk. Progress lives in two small files next to the archive, game.zip.hatchzip and game.zip.hatchzip-tail. Do not delete them while an extraction is unfinished: they hold the archive's index, which is no longer inside the archive itself.

To give up on an unfinished extraction instead, restore the archive:

hatchzip restore game.zip

What is left of the archive becomes a normal archive again, holding exactly the entries not yet extracted, which 7-Zip, WinRAR, unzip or tar open as usual. The files already extracted stay in the output folder. Restore only appends a new index, so it needs only the free space that index takes, and if it is interrupted, running it again finishes it.

hatchzip list game.7z    # what's inside, with sizes
hatchzip test game.7z    # read every file and verify its checksum, writing nothing

Plain tar has no checksums for file contents, so test on a tar only proves every file can be read in full.

HatchZip destroys the archive as it extracts, so it is built around never losing a file that has not safely landed on disk.

• Nothing changes until the plan fits. The space check runs before the first byte is touched, and it replays the whole extraction to find the moment that needs the most free space, not just the total.
• A file is written, flushed to the physical disk and renamed into place before the part of the archive holding it is cut off. A crash at any point leaves either the file or its bytes in the archive, never neither.
• Damage stops the run where it is. Every file's CRC-32 is checked as it is extracted. A damaged file is not written, and it stays in the archive along with everything before it.
• Hostile archives are refused before anything is deleted: paths that escape the output folder, entries that overlap each other, entries that appear twice. Setuid and setgid bits are never restored, and symlinks are created only after every regular file is written, so nothing can be written through one.

What it cannot protect against: if the download itself is damaged, the files before the damaged spot are extracted and the archive is partly gone. Use --verify-first when the archive can't be downloaded again. On btrfs or APFS, space held by a snapshot that includes the archive is not freed until the snapshot is deleted.

ZIP, 7z and tar all store their contents as a sequence of independent pieces: files for ZIP and tar, solid blocks for 7z. ZIP and 7z keep their index at the very end. HatchZip saves that index to a side file, then extracts pieces starting from the last one in the archive. After each batch it truncates the archive to where that batch began, which every operating system and filesystem supports; for a split archive, that means deleting whole volumes and shortening one. A .tar.gz is one compressed stream that can only be decoded from the front, so HatchZip extracts it front to back and punches holes in the compressed bytes it has used, at resume points it can restart from after a crash. The full design, including the exact order of writes that makes a crash safe, is in How it works.

HatchZip is licensed under either of Apache License, Version 2.0 or MIT license, at your option.

Bug reports, especially from filesystems and operating systems not covered by CI, are welcome. Read CONTRIBUTING.md first.

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in HatchZip by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.