Skip to content
Alpha — private preview. Expect rough edges and breaking changes between releases.

Windows via WSL2

There is no native Windows build of Agentarivm. The manager, and the tools it orchestrates — git, gh, Claude Code — are Unix programs, and WSL2 runs them as exactly that: a real Linux environment on a Windows machine.

  1. Install WSL2 with a Linux distribution (Ubuntu is fine):

    Terminal window
    wsl --install
  2. Inside WSL2, install the prerequisites — git, the GitHub CLI and Claude Code — and then Agentarivm itself:

    Terminal window
    curl -fsSL https://agentarivm.com/install.sh | sh

    The installer treats WSL2 as what it is: Linux x64.

  3. Check the machine:

    Terminal window
    agentarivm doctor

From here, everything in Getting started applies unchanged — WSL2 is the Linux platform.

  • Keep everything inside the Linux filesystem. Clone repositories under your WSL2 home (e.g. ~/git), not under /mnt/c/…. Git operations across the Windows filesystem boundary are dramatically slower, and worktrees belong on the Linux side.
  • An always-on manager needs an always-on WSL2. WSL2 shuts down when nothing keeps it alive. For a machine meant to serve sessions unattended, configure the systemd user service (Installation), enable lingering with loginctl enable-linger, and keep the WSL2 instance running — recent Windows builds support systemd in WSL2 out of the box, and wsl --exec from Task Scheduler at logon is a common way to keep the distribution up.
  • Reaching the manager from your phone means reaching the WSL2 VM. WSL2 has its own virtual network; forward the manager’s port from Windows to the VM (netsh interface portproxy, or WSL’s mirrored networking mode on recent builds) before setting server.publicUrl.

A dedicated Linux box or Mac remains the smoothest home for the manager; WSL2 is a supported way in, not the recommended end state for an always-on deployment.