Installation
Download the latest VGhostSetup.exe from the
Download page and run it.
The installer will place V-Ghost in %LocalAppData%\Programs\V-Ghost
and create a Start Menu shortcut.
V-Ghost requires the .NET 9 Desktop Runtime. If it is not already installed, the setup wizard will offer to download it automatically.
First Launch
On first launch V-Ghost will ask you to:
- Grant microphone permission — Windows will show a permission prompt, click Allow.
- Choose your default language (English is pre-selected).
- Choose your recognition engine — Vosk (recommended, offline) or Windows SAPI.
If you chose Vosk, the app will download the language model (~50–150 MB depending on language) on first use. An internet connection is only required for this initial download; afterwards everything runs offline.
Create a Profile
Profiles group voice commands together. You can have as many profiles as you like — for example one for a game, one for your code editor, one for general Windows control.
- Open the Profiles tab (second icon in the bottom bar).
- Tap the + button to create a new profile.
- Give it a name, e.g.
Visual Studio Code. - Optionally set a Confidence Threshold (0–1). The default of
0.75is a good starting point — lower it if commands aren't triggering. - Press Save.
Add Voice Commands
Inside a profile, each command maps a spoken phrase to one or more actions.
- With a profile open, tap Add Command.
- In the Phrase field, type the spoken trigger — e.g.
save. - Choose an Action Type:
- Key Press — simulate a key or modifier combo
- Mouse Click — left, right, or middle click at a target
- Process Launch — start an application or script
- Delay — pause for N milliseconds (useful in sequences)
- Configure the action. For Key Press, select the key and any modifier (
Ctrl,Shift,Alt,Win). - Press Save.
Example: "save" → Ctrl+S
Phrase: save
Action: Key Press
Key: S
Modifiers: Ctrl
Say "save" while the profile is active and V-Ghost sends Ctrl+S.
Macro sequences
You can chain multiple actions into a single command using a Macro Sequence. Add multiple steps (key presses, delays, etc.) in order — they execute in sequence when the phrase is heard.
Vosk vs Windows SAPI
V-Ghost supports two recognition backends. Both work fully — choose based on your preference:
| Vosk | Windows SAPI | |
|---|---|---|
| Internet required | Only for model download | Never |
| Accuracy | Excellent | Good |
| Languages | 29 (downloadable models) | Whatever Windows supports |
| Extra download | ~50–150 MB per language | None |
| Custom vocabulary | No | Yes (Windows Speech Training) |
Auto-Activate Profiles
A profile can be set to automatically activate when a specific application is in the foreground. This means your gaming profile switches on the moment you alt-tab into your game and switches back when you leave it.
- Open a profile and scroll to Auto-Activate.
- Enter the process name of the target app — e.g.
Codefor VS Code orEliteDangerous64for Elite Dangerous. - Save the profile. It will now activate and deactivate automatically.
Ctrl+Shift+Esc) → Details tab to find the exact executable name.
Settings Reference
Recognition Engine
Switch between Vosk and Windows SAPI globally, or override per profile.
Push-to-Talk
When enabled, recognition only runs while you hold down the configured hotkey. Useful if you don't want accidental triggers during normal conversation.
Overlay
A small on-screen indicator that shows the last recognized phrase and the active profile. Can be positioned in any corner of the screen (Top-Left, Top-Right, Bottom-Left, Bottom-Right) or disabled entirely.
Command Timeout
How long (in seconds) V-Ghost waits after a phrase is heard before discarding the match. Default: 30 seconds.