Run Discord in an Isolated Hyper-V Virtual Machine with Controlled Mic, Camera, and Screen Sharing Access

Overview

Warning: These instructions are AI generated and may contain inaccuracies or require adjustments based on your specific hardware, software versions, and use case. Please review each step carefully before executing.

This guide describes how to run Discord inside an isolated Windows virtual machine while still allowing controlled access to selected host resources such as microphone, audio output, webcams, and screen content. The goal is to keep Discord separated from your main operating system while still enabling the normal features people expect during calls or streaming.

Instead of running Discord directly on the host computer, the application runs inside a Hyper-V virtual machine. Hyper-V is Microsoft’s built-in virtualization platform for Windows. By placing Discord in a VM, the application cannot directly access your host desktop, files, or other applications. The only information Discord receives is what you intentionally provide.

The setup uses several technologies working together:

  • Hyper-V virtual machine to isolate Discord from the host system.
  • Enhanced Session Mode (RDP device redirection) to selectively allow devices such as microphones, audio playback, and cameras into the VM.
  • XSplit Broadcaster’s Virtual Camera to provide a controlled video feed from the host (for example your desktop or specific applications).
  • Optional GPU partitioning so the VM can use part of your graphics card for smoother video encoding and better performance.
  • A simple script or hotkey mechanism that allows you to quickly enable or disable microphone access to the VM.

This architecture provides several advantages:

  • Discord runs in a contained environment that cannot inspect or capture your host desktop directly.
  • You can share host applications or your screen through XSplit without giving Discord direct access to your system.
  • Access to sensitive devices such as microphones can be turned on or off intentionally.
  • Video encoding and screen sharing performance remain smooth through GPU acceleration.
  • The VM can be reset, paused, or destroyed without affecting your main Windows installation.

This configuration is commonly used by streamers, security researchers, and privacy-conscious users who want to limit the permissions granted to communication applications while still retaining full functionality for voice, video, and screen sharing.

The following instructions walk through the process step-by-step, beginning with enabling Hyper-V on Windows and finishing with configuring device redirection, virtual camera integration, GPU acceleration, and microphone control.

Guide

What you’re building (high level)

  • Discord runs inside a Hyper-V Windows VM (isolated).

  • The VM only gets:

    • camera feed = XSplit Virtual Camera (host-captured content)
    • optional microphone = redirected over Enhanced Session (RDP)
    • audio playback = redirected (optional)
    • Discord inside the VM cannot “see” your host desktop/apps directly; it only sees the VM + whatever you intentionally send (XSplit virtual camera and redirected devices).

1. Finish enabling Hyper-V (you are at RestartNeeded: True)

  • Do this now:

    • Re-run the command and answer Y, or just reboot from Start menu.

  • After reboot, confirm Hyper-V is enabled:

Get-WindowsOptionalFeature -Online -FeatureName Microsoft-Hyper-V

2. Make sure CPU virtualization is enabled in BIOS/UEFI (critical)

  • On reboot, enter BIOS/UEFI

  • Enable:

    • Intel: ‘Intel Virtualization Technology (VT-x)’
    • AMD: ‘SVM’ / ‘AMD-V’

  • Save and boot back into Windows

  • Quick check in Windows:

systeminfo | findstr /i "Virtualization"
  • You want to see: ‘Virtualization Enabled In Firmware: Yes’

3. Create the Discord VM (Hyper-V Manager)

  • Open: Hyper-V Manager
  • Action → New → Virtual Machine

  • Use these settings (good defaults for your hardware)

    • Generation 2
    • Memory: 8192 MB or 12288 MB (you can go higher later)
    • Networking: Default Switch (fine to start)
    • Virtual disk: 80 GB or more
    • Install Windows 11 in the VM (ISO)

    • Inside the VM:

    • Run Windows Update until fully updated

    • Install Discord

4. Turn on Enhanced Session Mode (this is what gives mic/cam/audio redirection)

  • On the host:

    • Hyper-V Manager → click your host name (left pane) → Hyper-V Settings…
    • Enhanced Session Mode Policy → Enable
    • Enhanced Session Mode → Enable
    • Microsoft doc for this workflow: (Microsoft Learn)

5. Connect to the VM with device redirection enabled (camera + optional mic)

  • In Hyper-V Manager:

    • Right-click the VM → Connect
    • In the VMConnect window, choose Enhanced Session and click ‘Show Options’

    • Enable:

    • Audio playback (to hear Discord)

    • Audio recording (this is the mic)
    • ‘Video capture devices’ (this is camera redirection)

    • Notes:

    • Camera/video capture redirection is standard RDP behavior and is exactly what Enhanced Session Mode uses under the hood. (Microsoft Learn)

6. Set up XSplit on the host to output your host screen as a ‘camera’

  • On the host:

    • Open XSplit Broadcaster
    • Add sources you want to share (host desktop capture, specific apps, etc.)
    • Enable XSplit Virtual Camera output (XSplit creates a virtual webcam device)
    • Goal: your host desktop/app scene becomes a webcam feed.

7. Make Discord in the VM use the XSplit Virtual Camera feed

  • In the VM:

    • Discord → User Settings → Voice & Video
    • Camera: select ‘XSplit Virtual Camera’ (it should appear via the redirected ‘Video capture devices’)
    • Turn on Video in a call

    • Result:

    • People in Discord see your host desktop/app scene (coming from XSplit), but Discord is still running inside the isolated VM.

8. Mic toggle with a hotkey (simple + reliable)

  • The cleanest toggle is: launch the VM session in two modes

    • ‘Mic ON’ session
    • ‘Mic OFF’ session

  • You switch by disconnecting and reconnecting (fast, predictable, no mystery states).

8A) Create two RDP connection profiles (on the host)

* Press Win+R → type ‘mstsc’ → Enter
* Click ‘Show Options’
* Local Resources tab:

  * For MIC ON profile:

    * Remote audio: ‘Play on this computer’
    * Remote audio recording: ‘Record from this computer’
    * Local devices and resources → More… → check ‘Video capture devices’
  * For MIC OFF profile:

    * Remote audio recording: ‘Do not record’
    * More… → check ‘Video capture devices’
* Save as:

  * ‘discord_vm_with_mic.rdp’
  * ‘discord_vm_no_mic.rdp’

* Edit each .rdp in Notepad and set:

  * Mic ON:

    * ‘audiocapturemode:i:1’
  * Mic OFF:

    * ‘audiocapturemode:i:0’
* (This property is commonly used for RDP audio capture control.) ([Microsoft Learn][3])

8C) Hotkeys (AutoHotkey v2)

* Install AutoHotkey v2 on the host
* Create a file ‘discord_vm_hotkeys.ahk’:

#Requires AutoHotkey v2.0

; Ctrl+Alt+M = connect with mic
^!m::Run 'mstsc.exe "' A_ScriptDir '\discord_vm_with_mic.rdp"'

; Ctrl+Alt+N = connect no mic
^!n::Run 'mstsc.exe "' A_ScriptDir '\discord_vm_no_mic.rdp"'

- Run the script
- Now Ctrl+Alt+M or Ctrl+Alt+N is your “toggle” (by reconnecting)

9. GPU acceleration in the VM (GPU partitioning / GPU-P)

  • This is the part that sometimes varies by driver/GPU support, but your RTX 5090 is exactly the type of card people use for this.

  • Official background + steps are documented by Microsoft. (Microsoft Learn)

9A) Add the GPU partition adapter (host PowerShell as Admin)

* Stop the VM first:

Stop-VM -Name 'DiscordVM' -Force

- Add GPU partition adapter:

Add-VMGpuPartitionAdapter -VMName 'DiscordVM'

- Give the VM enough MMIO space (common requirement):

Set-VM -Name 'DiscordVM' -GuestControlledCacheTypes $true -LowMemoryMappedIoSpace 1GB -HighMemoryMappedIoSpace 8GB

- Start VM:

Start-VM -Name 'DiscordVM'

9B) Verify in the VM

  • Open ‘dxdiag’ → Display tab
  • Device Manager → Display adapters
  • If you still see only ‘Microsoft Basic Display Adapter’, GPU-P didn’t attach yet.

9C) If GPU-P doesn’t attach on first try (most common fixes)

  • Fully update NVIDIA drivers on the host
  • Fully update Windows on host + guest
  • Power off the VM (not just restart), then start again
  • If needed, remove and re-add the GPU partition adapter:
Stop-VM -Name 'DiscordVM' -Force
Remove-VMGpuPartitionAdapter -VMName 'DiscordVM'
Add-VMGpuPartitionAdapter -VMName 'DiscordVM'
Start-VM -Name 'DiscordVM'

10. Confirm the isolation behavior you want

  • Discord in the VM can’t capture your host apps directly.
  • The only host visuals it can share are what you intentionally feed it (XSplit Virtual Camera).
  • The only mic access it gets is what you allow via the RDP profile / hotkey (mic on/off).

  • Quick notes tailored to your setup

  • You have multiple webcams (StreamCam/document cams). Those can be redirected as ‘Video capture devices’ in Enhanced Session.

  • If you want a “hard off” for mic beyond redirection:

    • also mute the mic inside the VM (Discord input device disabled), and/or disable the audio capture device in Device Manager inside the VM.

If you tell me the exact VM name you created (or what you want it to be), I’ll paste the GPU-P commands with that name already filled in, plus a pair of ready-to-use .rdp file templates (mic on/off) that you can drop into a folder and bind to hotkeys.