> ## Documentation Index
> Fetch the complete documentation index at: https://docs.scoutvoice.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Dictation

> How voice-to-text dictation works in Scout.

Dictation is Scout's core feature. Hold a hotkey, speak, and Scout types what you say into whatever app has focus.

When [context awareness](/features/context-awareness) is on, Scout also uses surrounding text from the focused app to improve recognition, correction, and insertion behavior.

## Push-to-talk

The default dictation mode. Hold your dictation hotkey, speak, and release to inject text.

* **Windows:** Hold **Ctrl+Shift** to dictate. Release to inject.
* **macOS:** Hold **Fn** to dictate. Release to inject.

<Steps>
  <Step title="Place your cursor">
    Click into any text field in any app.
  </Step>

  <Step title="Hold the dictation hotkey">
    Scout starts streaming audio to the speech engine immediately.
  </Step>

  <Step title="Speak naturally">
    Talk at a normal pace. You can say [voice commands](/voice-commands/overview) inline — they're detected and converted in real time.
  </Step>

  <Step title="Release to inject">
    Scout finalizes the text — applying [dictionary](/features/dictionary), [quick phrases](/features/quick-phrases), and [app style](/features/app-styles) — then injects it at your cursor.
  </Step>
</Steps>

## Hands-free mode

For longer dictation, use hands-free mode so you don't have to hold a key.

* **Windows:** Press **Ctrl+Alt+Space** to start. Press again to stop.
* **macOS:** Press **Fn+Space** to start. Press again to stop.

Hands-free sessions have two automatic safeguards:

* **15-minute maximum** — Sessions automatically end after 15 minutes.
* **90-second silence stop** — If Scout detects no speech for 90 seconds, it stops the session.

Text is accumulated during the session and injected when the session ends.

## How text injection works

Scout uses a clipboard-based injection method:

1. Your current clipboard text is saved
2. The transcribed text is copied to the clipboard
3. Scout simulates a paste keystroke (Ctrl+V on Windows, Cmd+V on macOS)
4. Your original clipboard text is restored

<Note>
  The clipboard save/restore preserves text content. If you have non-text content (images, files) on the clipboard, it may not be restored after injection.
</Note>

## Processing pipeline

Between your voice and the final text, Scout runs several processing steps on your machine:

1. **Context awareness** — If enabled, Scout reads the focused field and nearby focused-app context to improve recognition and corrections. See [Context Awareness](/features/context-awareness).
2. **Voice commands** — Punctuation, symbols, line breaks, and formatting commands are detected and converted. See [Voice Commands](/voice-commands/overview).
3. **Auto-list formatting** — Digit-based numbered lists from the speech engine are formatted into clean list items.
4. **Vocabulary corrections** — Words in your [dictionary](/features/dictionary) are matched and replaced using fuzzy matching.
5. **Quick phrases** — Trigger phrases are expanded to their full replacements. See [Quick Phrases](/features/quick-phrases).
6. **App style** — If you have an [app style](/features/app-styles) configured for the current app, it's applied (unless a quick phrase was expanded).
7. **Text injection** — The final text is injected at your cursor.

<Tip>
  If something doesn't look right, say **"undo that"** — Scout sends the undo shortcut (Ctrl+Z / Cmd+Z) to the focused app. Note that this undoes whatever is on the app's undo stack, which may include non-Scout actions if you've typed or edited since dictating.
</Tip>