What is Debark? — Debark Docs

Debark prepares software for an offline Debian or Ubuntu machine. It downloads the packages you ask for, plus the dependencies that machine needs, into a folder you can copy over.

The basic workflow

Choose a baseline OS, or capture a snapshot of the offline machine’s installed packages and apt settings.
Use that target on an online computer to download your packages and their missing dependencies.
Copy the resulting bundle to the offline machine and install it with the CLI.

A baseline OS lets you start without a snapshot by choosing a stock Debian or Ubuntu installation. Use a snapshot when you need the machine’s actual state. Either way, dependency planning uses the selected target’s package set.

CLI and desktop app

The CLI handles the whole workflow: snapshots, downloads, verification, and installation. Its executable is currently named debark.

Run debark build --interactive for guided questions, or pass the target and packages directly as command-line options.

The desktop app runs on the online Linux computer. It adds a searchable package catalogue and guides you through building and exporting a bundle.

Both are free and open source under Apache-2.0. There’s no account to create.

What’s in a bundle?

A bundle contains .deb packages, an apt repository index, a list of exact package versions, and checksums. You can sign it with your own key so the offline machine can verify its source and contents before installing.

The folder also includes a README.txt summary. See the bundle format reference for the full file list.

Requirements and limits

The offline machine must use Debian or Ubuntu apt packages. Downloads happen on the online computer, using matching apt tools or a Docker / Podman container.

Snap and Flatpak apps aren’t included. Some .deb packages need network access during installation; check the bundle before transferring it.

The project is pre-1.0, with CLI and desktop downloads and source builds available from the public repository. Check known limitations for validation coverage. Feature requests, questions, and bug reports are welcome.