AudioSwitchPS

Switch default audio devices (output + input) using profiles — keyboard shortcuts, a system tray icon, autostart. No external dependencies, no administrator rights required.

Same architecture as NetSetPS: a precompiled DLL instead of runtime Add-Type, a shortcut with -WindowStyle Minimized + HideConsole() inside the script, a NotifyIcon with a manual MouseUp handler.

Install

powershell -ExecutionPolicy Bypass -File .\Install.ps1

What it does:

• copies the scripts to %LOCALAPPDATA%\AudioSwitchPS
• compiles AudioSwitchPS.Native.dll (csc.exe from .NET Framework 4.x — present on every Windows install)
• creates a Start Menu shortcut and a Startup shortcut
• launches the app

Switches: -NoAutostart, -NoLaunch, -Force (rebuild the DLL), -InstallPath.

Uninstall: .\Uninstall.ps1 [-RemoveConfig]

How it works

Layer	Mechanism
Device list, volume, mute	IMMDeviceEnumerator / IAudioEndpointVolume (Core Audio API)
Changing the default device	IPolicyConfig::SetDefaultEndpoint — undocumented COM, the only way to do it without the GUI
Global hotkeys	RegisterHotKey + an invisible NativeWindow catching WM_HOTKEY
Autostart	a .lnk shortcut in shell:startup (per-user, no Task Scheduler, no UAC)

Roles set on switch: eConsole + eMultimedia always, eCommunications when SetCommunicationsRole = true.

Autostart — no prompt is expected

Toggling "Run at Windows startup" in the tray menu does not show a UAC prompt — creating a shortcut in the per-user Startup folder doesn't need elevation, so there's nothing to confirm. Feedback comes as a MessageBox confirming whether the shortcut was actually created/removed (an earlier version relied on a balloon tip, which Windows 10/11 frequently suppresses — Focus Assist, notification settings, or simply firing a second balloon too soon — making a successful toggle look like nothing happened).

To verify manually: open shell:startup (Win+R) and check for AudioSwitchPS.lnk, or check the log at %APPDATA%\AudioSwitchPS\audioswitchps.log for an Autostart enabled: / Autostart disabled. line.

Configuration

%APPDATA%\AudioSwitchPS\audioprofiles.json — created on first run from the current devices. Full-field template: audioprofiles.example.json.

{
  "Settings": { "SetCommunicationsRole": true, "ShowBalloon": true, "BalloonTimeoutMs": 1500 },
  "Profiles": [
    {
      "Name": "Headset",
      "Hotkey": "Ctrl+Alt+1",
      "Output": { "Name": "*Jabra*", "Id": "{0.0.0.00000000}...", "Volume": 45, "Mute": false },
      "Input":  { "Name": "*Jabra*" }
    }
  ]
}

• Output / Input — both sections optional (a profile can change output only).
• Volume (0–100) and Mute — optional, skipped if the field is absent.
• Device matching order: Id first, then Name with a wildcard, then a name fragment. Worth keeping both — Id is fast and unambiguous, Name saves the profile after a USB device is moved to a different port (which changes its Id).

Hotkey syntax

Ctrl / Alt / Shift / Win + a key, e.g. Ctrl+Alt+1, Ctrl+Shift+F9, Win+Alt+Num1, Ctrl+Alt+Oemtilde.

A hotkey already in use elsewhere doesn't crash the app — you get a balloon listing the conflicts, and a WARN line in the log.

Appearance

The tray menu and the "Save current as profile" dialog use a dark steel/slate palette (AudioSwitchPS.Native.DarkColorTable) with subtly rounded corners (ApplyRoundedCorners — DWM corner preference on Windows 11, a GDI region fallback on Windows 10). Disabled header lines (current output/input) use a dimmed but still legible gray rather than the renderer's default near-invisible disabled color. Standard MessageBox dialogs (autostart confirmation, "already running") stay OS-native and aren't restyled — theming those would require a custom owner-drawn dialog, which wasn't worth the added surface area here.

Menu (left or right click on the icon)

Current output/input · profile list (checkmark on the active one) · quick single-device selection with volume · Save current as profile · edit the JSON in Notepad with auto-reload on save · autostart toggle · Sound settings / Volume mixer · Exit.

CLI mode

.\AudioSwitchPS.ps1 -List                    # devices + profiles, asterisk = default
.\AudioSwitchPS.ps1 -Apply "Headset"         # matches by name, partial match too
.\AudioSwitchPS.ps1 -SetOutput "*Realtek*"
.\AudioSwitchPS.ps1 -SetInput  "*Jabra*"

CLI mode exits immediately and never creates a tray icon — good for calling from another script, a desktop shortcut, or a PowerShell profile function.

Log

%APPDATA%\AudioSwitchPS\audioswitchps.log, rotated at 1 MB. Levels: INFO, WARN, ERROR, PERF (profile apply / menu build time in ms).

Notes

• Only devices in the active state (DEVICE_STATE_ACTIVE) are listed — disconnected or disabled devices won't show up.
• If an AV product flags the compilation step during install: that's a one-time csc.exe run, it never happens at runtime. If needed, build the DLL once on a reference machine and distribute the finished file.
• The same CN=NetSetPS Self-Signed certificate can be reused to sign the code — Set-AuthenticodeSignature on the *.ps1 files and the DLL.
• Both PS 5.1 and PS 7 work; the shortcuts deliberately target 5.1 (System32\WindowsPowerShell\v1.0\powershell.exe) since it's always present.