Skip to main content

Wake Word Feature

Wake word detection allows you to activate SlasshyWispr’s assistant mode using a voice phrase instead of pressing a hotkey. Simply say your wake phrase (like “Hey Lily”) to start a conversation with your voice assistant. This feature enables:
  • Hands-free assistant activation
  • Natural conversation flow
  • Quick access while typing or working
  • Always-listening assistant mode when enabled
Wake word detection is part of SlasshyWispr’s command mode and assistant features. When enabled, the app listens for your wake phrase in the background.

Configuration Settings

Wake word functionality is controlled by two key settings:

Wake Word Enabled

Setting: wakeWordEnabled (boolean) Default: false (disabled) Enables or disables wake word detection. When turned on, SlasshyWispr listens for your wake phrase and automatically enters assistant mode when detected.
1

Open General Settings

Navigate to Settings > General in SlasshyWispr.
2

Enable Wake Word

Toggle the Wake Word Enabled option to turn on wake word detection.
3

Test Activation

Say your wake phrase (e.g., “Hey Lily”) to test the activation.

Assistant Name

Setting: assistantName (string) Default: "Lily" The name your assistant responds to. This forms the wake phrase pattern (e.g., “Hey Lily”, “OK Lily”).
1

Choose a Name

Pick a distinctive name that’s easy to pronounce and unlikely to trigger accidentally. The default is “Lily”.
2

Update Assistant Name

In Settings > General, change the Assistant Name field to your preferred name.
3

Use Natural Phrases

Your wake phrase typically follows patterns like:
  • “Hey [Name]”
  • “OK [Name]”
  • “[Name]“

How Wake Word Detection Works

When wake word detection is enabled, SlasshyWispr:
1

Background Listening

Continuously listens to your microphone in the background for the wake phrase.
2

Wake Phrase Recognition

When it detects a phrase matching your wake word pattern (“Hey [AssistantName]”), it activates.
3

Assistant Mode Activation

Automatically switches to assistant mode and begins recording your command or question.
4

Processing & Response

Processes your speech, generates an AI response, and speaks the answer using TTS (text-to-speech).
Wake word detection requires continuous microphone access. Ensure your microphone is properly configured in Settings > General and that SlasshyWispr has necessary permissions.

Customizing Wake Phrases

While SlasshyWispr uses your configured assistantName, you can optimize your wake phrase experience:

Choosing an Effective Name

Choose names with 2-3 syllables that are phonetically distinct. Names like “Lily”, “Jarvis”, “Atlas” work well.Good examples: Lily, Jarvis, Atlas, Nova, EchoAvoid: Common words, single syllables, names that sound like common phrases
Don’t use names that sound like words you frequently say (“Hey”, “OK”, common names of people around you).This reduces false activations during normal conversation.
Test your chosen name in your actual working environment. Say it in different tones and volumes to ensure consistent detection.

Wake Phrase Patterns

SlasshyWispr recognizes wake phrases that include your assistant name. Common patterns:
  • “Hey [Name]”: Most natural and commonly used (e.g., “Hey Lily”)
  • “OK [Name]”: Alternative activation phrase (e.g., “OK Lily”)
  • “[Name]”: Direct name call may work depending on implementation
Speak clearly and at normal volume. Shouting or whispering may reduce detection accuracy.

Tips for Reliable Activation

Microphone Setup

1

Select Quality Microphone

Use a good quality microphone for best detection. USB microphones or headset mics typically work better than built-in laptop mics.
2

Check Microphone Settings

In Settings > General, verify the correct microphone device is selected under Microphone Device.
3

Test Audio Levels

Ensure your microphone picks up your voice clearly without distortion or excessive background noise.

Environment Optimization

Reduce Background Noise: Wake word detection works best in quiet environments. Reduce background music, conversations, or ambient noise when possible.
Maintain a consistent distance from your microphone. For desktop mics, 6-12 inches typically works well.
Speak the wake phrase clearly and naturally. Don’t rush or mumble the assistant name.
After saying the wake phrase, pause briefly (about 1 second) before stating your command or question. This gives the system time to fully activate.

Troubleshooting Detection Issues

Problem: Assistant activates when you don’t want it to.Solutions:
  • Choose a more distinctive assistant name
  • Reduce microphone sensitivity if your OS/hardware allows
  • Disable wake word when not needed
  • Check for background audio (TV, music) that might contain similar sounds
Problem: Wake word doesn’t trigger when you say it.Solutions:
  • Speak louder and more clearly
  • Check microphone is selected correctly in settings
  • Test microphone in system settings to ensure it’s working
  • Verify wake word is enabled (wakeWordEnabled: true)
  • Try different wake phrase patterns (“Hey Lily” vs “OK Lily”)
Problem: Long delay between wake word and activation.Solutions:
  • Check CPU usage - background processes may slow detection
  • Ensure local STT models are warmed up if using offline mode
  • Close unnecessary applications to free system resources

Command Mode vs Wake Word

SlasshyWispr offers multiple ways to activate assistant mode:
MethodActivationUse Case
Wake WordVoice phrase (“Hey Lily”)Hands-free, always available
Command HotkeyKeyboard shortcut (default: Ctrl+Shift+Space)Quick activation while typing
Push-to-TalkHold hotkey (default: Ctrl+Space)Precise control of recording
You can use wake word and hotkeys together! Enable wake word for hands-free access, but still use hotkeys when you prefer keyboard control.

When to Use Wake Word

  • During hands-free work (cooking, exercising, away from keyboard)
  • While typing or using the mouse
  • For quick queries without interrupting workflow
  • When you want natural, conversational assistant access

When to Use Hotkeys Instead

  • In noisy environments where wake word detection is unreliable
  • When you need precise control over recording start/stop
  • During meetings or calls (to avoid accidental activation)
  • When wake word false positives are problematic

Privacy Considerations

Microphone Access: When wake word detection is enabled, SlasshyWispr requires continuous access to your microphone to listen for the wake phrase.If you’re concerned about always-on listening:
  • Use hotkeys instead of wake word
  • Disable wake word when not needed
  • Check that you’re using local STT models if you want processing to stay on-device
Local Processing: If you’re using local STT models (offline mode), wake word detection happens entirely on your device with no data sent to external servers.

Example Configuration

Full Wake Word Setup

{
  "wakeWordEnabled": true,
  "assistantName": "Lily",
  "commandMode": true,
  "microphoneDeviceId": "default",
  "sttRuntimeMode": "local",
  "localSttModel": "nvidia/parakeet-tdt-0.6b-v3"
}
This configuration:
  • Enables wake word detection
  • Sets assistant name to “Lily” (say “Hey Lily” to activate)
  • Uses local STT for privacy
  • Keeps command mode enabled for assistant features

Alternative: Hotkey-Only Setup

{
  "wakeWordEnabled": false,
  "commandHotkey": "Ctrl+Shift+Space",
  "pushToTalkHotkey": "Ctrl+Space"
}
This configuration:
  • Disables wake word
  • Uses keyboard shortcuts only
  • No continuous microphone listening
Start with wake word disabled, learn the hotkeys, then enable wake word once you’re comfortable with the assistant features.

Build docs developers (and LLMs) love