Ardela Voice for desktop
Stream dictation and insert reviewed text into a compatible active field.
Ardela Voice for desktop is the quick direct-insertion surface within Voice. It streams your dictation, prepares the text, and attempts to insert it into the compatible text field you were using.
Desktop dictation does not create an Ardela web Recording. It does not feed the browser timeline, template-generation, client, matter, or Queue workflow.
Platforms
Ardela Voice is available for Windows and macOS. Ask your Workspace administrator or Ardela support for the approved installer.
Dictate into an active field
- Place the cursor in a compatible text field in the application where you are working.
- Use the configured global hotkey.
- Speak while the Voice widget shows the live preview.
- Stop the recording using the selected trigger mode.
- Ardela applies the selected clean-up mode and attempts to insert the result into the captured field.
- Review the inserted wording before saving, sending, or filing it.
Desktop is the only Voice surface documented as real-time. Browser and native-mobile recordings are transcribed in batch after Finalise.
Trigger modes and default hotkeys
Choose either:
- Hold to record: hold the shortcut while speaking and release to stop.
- Press to start/stop: press once to begin and again to stop.
The defaults are:
| Platform | Default |
|---|---|
| Windows | Alt |
| macOS | Ctrl+Shift |
Change the shortcut and trigger mode in desktop settings if they conflict with another application.
Compatible text fields and permissions
Insertion depends on the target application's text controls and operating-system permissions:
- macOS requires Accessibility permission for the hotkey and text-control path;
- some browser editors, document canvases, terminals, and other complex fields use a verified clipboard-paste fallback;
- a Windows application running with elevated privileges may require Ardela Voice to run with matching privileges;
- password fields are not read from or written to.
If the target refuses the text, Ardela leaves the result on the clipboard so you can paste it manually.
Clean-up and correction
Desktop Voice supports a live preview, custom Workspace vocabulary, verbatim output, and AI-assisted clean-up or rewrite modes. These capabilities are specific to desktop and must not be assumed for browser or mobile batch transcription.
After insertion, the correction bubble can offer:
- Undo to revert the insertion;
- Retry to run the cleaner again;
- Verbatim to replace it with the uncleaned transcript.
Protected local history
Desktop history and replay audio are scoped to the signed-in user and Workspace, encrypted locally with keys protected by the operating system, and automatically deleted after 7 days.
A single desktop recording stops at 20 minutes or before its local audio buffer exceeds 48 MiB.
Human review
Real-time transcription and clean-up can omit or change important content. Check names, dates, amounts, undertakings, advice, and legal meaning in the target application before the text leaves your control.