Voice Assistant

CueCapture’s Voice Assistant lets you add tasks hands-free. Just speak naturally, and CueCapture will create tasks with dates, priorities, and more.

Enabling Voice

1. Open Settings (⚙️)
2. Go to Voice Assistant
3. Enable Voice Input
4. Grant microphone permission when prompted

Quick Start

1. Press V or click the microphone icon
2. Say your task naturally:
   • “Call mom tomorrow”
   • “Submit report by Friday”
   • “High priority: review budget”
3. CueCapture creates the task automatically

Voice Modes

Browser Speech

Uses your browser’s built-in speech recognition.

Pros:

• No setup required
• Works offline (sometimes)
• Free

Cons:

• Accuracy varies by browser
• Limited language support
• Requires browser permissions

Whisper (Local)

Uses OpenAI’s Whisper model running locally.

Pros:

• Excellent accuracy
• Works completely offline
• Privacy-focused

Cons:

• Requires model download (~500MB)
• Uses more system resources
• Initial processing delay

Deepgram (Cloud)

Uses Deepgram’s cloud speech API.

Pros:

• Best accuracy
• Very fast
• Many languages

Cons:

• Requires API key
• Needs internet connection
• May have costs at high usage

Configure in Settings → Voice Assistant → Speech Mode.

Natural Language

CueCapture understands natural language for:

Dates

• “tomorrow”
• “next Tuesday”
• “in 3 days”
• “December 15th”
• “end of month”
• “next week”

Times

• “at 2pm”
• “by 5:30”
• “morning” (→ 9am)
• “afternoon” (→ 2pm)
• “evening” (→ 6pm)

Priority

• “high priority”
• “urgent”
• “important”
• “low priority”
• “when I get to it”

Projects

• “for Project Alpha”
• “in marketing project”
• “add to Q4 Planning”

Examples

You Say	Task Created
”Call dentist tomorrow at 2pm”	Call dentist • Due: Tomorrow 2:00 PM
”Urgent: fix login bug”	fix login bug • Priority: High
”Write report for Q4 project by Friday”	Write report • Project: Q4 • Due: Friday
”Low priority: organize desk”	organize desk • Priority: Low

Voice Commands

Beyond creating tasks, you can use voice commands:

Task Management

• “Show my tasks”
• “What’s due today?”
• “Show high priority tasks”
• “Complete [task name]”
• “Delete [task name]“

Navigation

• “Go to projects”
• “Open settings”
• “Show matrix view”
• “Filter by [project name]“

Timer

• “Start timer on [task name]”
• “Stop timer”
• “How much time today?”

Miscellaneous

• “What time is it?”
• “Help” - shows available commands

Note

Command recognition works best with Whisper or Deepgram modes. Browser speech may have limited command support.

Microphone Setup

Select Microphone

1. Settings → Voice Assistant → Microphone
2. Select your preferred microphone
3. Test with the “Test” button

Troubleshooting Audio

• Ensure microphone permissions are granted
• Check system sound settings
• Try a different microphone
• Speak clearly and at normal volume

Push-to-Talk vs Auto-Listen

Push-to-Talk (Default)

Press and hold V or click mic to record.

• More control
• No accidental activations
• Works in noisy environments

Auto-Listen

CueCapture listens when widget is focused.

• Hands-free operation
• Uses wake word (optional)
• May pick up background noise

Configure in Settings → Voice Assistant → Input Mode.

Wake Word

When Auto-Listen is enabled, you can use a wake word:

1. Enable Auto-Listen mode
2. Set your wake word (default: “Hey CueCapture”)
3. Say wake word to activate
4. Speak your command
5. CueCapture processes after silence

Custom Wake Words

• “Hey CueCapture”
• “Okay Tasks”
• “Computer” (for Star Trek fans)

Tip

Custom wake words require Whisper or Deepgram mode for reliable detection.

Keyboard Shortcuts

Shortcut	Action
V	Push-to-talk (hold to speak)
Shift + V	Toggle auto-listen
Ctrl/Cmd + V	Voice settings
Esc	Cancel voice input

Voice Feedback

CueCapture can respond with audio:

Text-to-Speech

Enable spoken confirmations:

• “Task created: Call dentist tomorrow”
• “Timer started on Review budget”
• “3 tasks due today”

Voice Selection

Choose from available system voices:

1. Settings → Voice Assistant → Voice
2. Select from dropdown
3. Preview with “Test” button

Greeting Style

Configure how CueCapture greets you:

• Simple: “Ready.”
• Summary: “Good morning! You have 5 tasks due today.”

Privacy

Browser Speech

• Audio processed by browser (Google/Apple/Microsoft)
• May send audio to cloud for processing
• Check browser privacy settings

Whisper (Local)

• All processing happens on your device
• No audio leaves your computer
• Completely private

Deepgram

• Audio sent to Deepgram servers
• Subject to Deepgram’s privacy policy
• API key required

Tips

Speak Naturally

Don’t use robotic phrasing. Natural speech works best:

• ✅ “Remind me to call mom tomorrow afternoon”
• ❌ “Add task call mom due date tomorrow time afternoon”

Be Specific

Include relevant details:

• ✅ “Urgent: review contract for Client X by end of day”
• ❌ “Review thing”

Use Projects

Include project names to auto-assign:

• “Add to website redesign: update homepage copy”

Corrections

If recognition is wrong:

• Say “Cancel” to discard
• Say “Edit that” to modify
• Or just create and edit manually

Troubleshooting

Voice Not Working

1. Check microphone permissions
2. Verify mic is selected in settings
3. Try a different speech mode
4. Check system audio settings

Poor Recognition

• Speak closer to microphone
• Reduce background noise
• Try Whisper or Deepgram mode
• Speak at normal pace

Commands Not Recognized

• Voice commands work best with Whisper/Deepgram
• Try phrasing differently
• Use explicit task creation: “Add task…”

Slow Processing

• Whisper mode requires model download
• First use may be slower
• Ensure adequate system resources
• Try Deepgram for faster results

Next Steps

• Task Management - Create tasks other ways
• Keyboard Shortcuts - Alternative quick input
• Settings - General troubleshooting