Skip to main content

​
Voice Dictation

SlasshyWispr’s voice dictation feature converts your speech into clean, accurate text with minimal friction. Activate it with a hotkey, speak naturally, and get instant transcription.

​
How Voice Dictation Works

Voice dictation is the primary mode for converting speech to text without AI processing:
  1. Activate using your configured hotkey (default: Ctrl+Space)
  2. Record your speech using single-tap or push-to-talk mode
  3. Transcribe audio through online or offline STT models
  4. Output text to clipboard or directly paste into your active application
Voice dictation focuses on accurate transcription. For AI-powered responses and context-aware rewriting, use Assistant Mode.

​
Capture Modes

SlasshyWispr supports two capture modes for recording audio:
Push-to-Talk (default) requires you to hold down the hotkey while speaking.
  • Hold the hotkey to record
  • Release to stop recording and process
  • Maximum recording time: 45 seconds
  • Ideal for precise control over recording duration
type CaptureMode = "push-to-talk" | "single-tap";
const DEFAULT_CAPTURE_MODE: CaptureMode = "push-to-talk";

​
Hotkey Activation

Dictation is triggered via global hotkeys that work across all applications:
HotkeyDefaultPurpose
Dictation HotkeyCtrl+SpaceActivates voice dictation mode
Command HotkeyCtrl+Shift+SpaceActivates assistant mode
Hotkeys are fully customizable in Settings > General. You can configure modifier keys (Ctrl, Shift, Alt, Meta) and the trigger key.

​
Recording Workflow

The dictation workflow follows these stages: 
type Stage = "idle" | "recording" | "processing" | "speaking" | "error";

​
Workflow Steps

  1. Idle - Application is ready and listening for hotkey activation
  2. Recording - Microphone is active and capturing audio
  3. Processing - Audio is being transcribed via STT model
  4. Complete - Transcript is ready and delivered to clipboard/paste
  5. Error - If transcription fails, an error message is displayed
Recordings are limited to 45 seconds (MAX_RECORDING_MS = 45_000) to ensure responsive processing.

​
Clipboard Integration

Transcribed text can be delivered via clipboard for flexible use:
  • Copy to Clipboard - Text is copied and ready to paste anywhere
  • Auto-Paste - Text is automatically pasted into the active application
  • Manual Control - Choose when and where to paste the transcription
These options are controlled by the following settings: 
interface PersistedSettings {
  autoPasteDictation: boolean;  // Auto-paste after dictation
  copyToClipboard: boolean;     // Copy to clipboard
  // ... other settings
}
Enable both Copy to Clipboard and Auto-Paste to get the best of both worlds: immediate pasting with a clipboard backup.

​
Auto-Paste Option

The auto-paste feature automatically inserts transcribed text into your active application:

​
How Auto-Paste Works

  1. You complete a dictation
  2. Text is transcribed via STT
  3. SlasshyWispr simulates keyboard input
  4. Text appears at your cursor position

​
Auto-Paste Settings

  • Enabled - Text is automatically pasted after transcription
  • Disabled - Text is only copied to clipboard (manual paste with Ctrl+V)
Auto-paste is configured in Settings > General under the autoPasteDictation setting.

​
When to Use Manual vs Auto-Paste

ScenarioRecommended Setting
Writing in text editorsAuto-paste enabled
Sensitive/password fieldsAuto-paste disabled
Reviewing before insertingAuto-paste disabled
Fast drafting workflowAuto-paste enabled

​
Text Enhancement Options

SlasshyWispr includes intelligent text processing features: 
interface PersistedSettings {
  backtrackCorrection: boolean;  // Smart error correction
  removeFillers: boolean;        // Remove "um", "uh", etc.
  autoPunctuation: boolean;      // Automatic punctuation
  numberedLists: boolean;        // Format numbered lists
}
  • Backtrack Correction - Automatically fixes common speech recognition errors
  • Remove Fillers - Strips out verbal fillers like “um”, “uh”, “like”
  • Auto Punctuation - Adds periods, commas, and other punctuation intelligently
  • Numbered Lists - Formats spoken numbers into proper list structures
Enable Remove Fillers and Auto Punctuation for cleaner, more professional dictation output.

​
Best Practices

  1. Speak Clearly - Natural pace with clear pronunciation works best
  2. Minimize Background Noise - Use a quality microphone in a quiet environment
  3. Use Wake Words - For assistant features, enable wake word detection
  4. Choose the Right Mode - Use dictation for transcription, assistant mode for AI responses
  5. Configure Hotkeys - Set hotkeys that don’t conflict with other applications

Build docs developers (and LLMs) love

Get started for freeTalk to us