CS2 case farming software.

What this software does.

MonkePanel's job is to give you back the hours you would otherwise spend logging in, launching CS2, walking around for a session, and collecting drops on thirty accounts every week. It replaces that loop with three coordinated jobs that run on autopilot on your Windows host.

Windows-native, single executable.

MonkePanel ships as a single signed Windows .exe that talks directly to Hyper-V. No Docker layer, no WSL passthrough, no Electron wrapper around a browser, no cloud agent calling home. The installer drops a native binary built with MSVC, registers a Windows service for background work, and uses the native window manager for the UI. Every VM is created against the Hyper-V WMI API at kernel level, which is the same surface Microsoft uses for its own virtualization tooling.

The reason for shipping as a native app rather than a browser tool or a SaaS is straightforward: Hyper-V GPU partitioning, low-latency input injection into a guest, and reading per-VM CPU and RAM counters all require API access that a browser cannot give you. Running close to the metal means lower overhead per VM, faster start and stop, and no extra network hop between the UI and the work. The software updates itself over a signed channel; you do not have to rebuild anything when CS2 patches. Read more in the Hyper-V setup guide.

System requirements.

The software itself is light. The Hyper-V guests it manages are not, because each one runs a full CS2 client. The numbers below are the floor, not the recommendation. A serious farm host runs above this. The full step-by-step is in the Hyper-V setup guide, including BIOS toggles and GPU partitioning notes.

• Windows 10 or 11 Pro, Enterprise, or Education

  Home edition does not include Hyper-V and cannot run the software. Read Microsoft's Hyper-V overview for confirmation.

• CPU with hardware virtualization

  Intel VT-x with VT-d, or AMD-V with AMD-Vi, enabled in BIOS. Without it, Hyper-V will not start a single guest.

• Discrete GPU with partitioning support

  NVIDIA GTX 1060 or AMD RX 480 or newer. Integrated graphics will not partition well enough to run CS2 inside a guest at usable frame rates.

• ~6 GB RAM per running VM

  Rough sizing for a CS2 guest. The host itself wants a few more gigabytes free for the Windows shell and the panel.

Install and first run.

Installation is intentionally boring. The alpha build arrives as a single signed installer. You run it, accept the UAC prompt, and the software lands in Program Files with a Start menu shortcut and a background service for the VM lifecycle. On first launch the app asks for one thing: the address of your local Hyper-V host. For most users that is the same machine; in a multi-PC farm it can be a different Windows host on the same LAN.

Once it has talked to Hyper-V the software walks you through adding your first Steam account. You drop in your credentials and import the maFile for Steam Guard, and the panel provisions a fresh VM with a fresh hardware fingerprint, installs CS2 inside it, and runs a one-time login. From that point the account is part of the farm and the walkbot takes over. We do not publish a public download link during the alpha; the build is gated behind a short chat. DM @monkecs on Telegram to apply. New testers should also read the getting started guide before first run.

Multi-account architecture.

One Windows host, many Hyper-V guests, one Steam account per guest. That is the entire mental model. The software you install on the host is the conductor. Each VM is a full Windows install with a partitioned slice of the host GPU and a distinct hardware fingerprint: its own CPU model string, its own MAC address, its own disk serial, its own GPU device ID. Steam sees a separate machine at every login, which is the entire point.

The host runs the panel UI you click and the background service that watches the guests. Inside each guest CS2 launches normally and the walkbot drives the player. There is no shared Steam install between VMs and no shared profile state, so a ban on one account does not cascade into the rest of the cluster. For a different cut at the same architecture, see the page about the underlying farming bot; for the management surface itself see the panel UI.