Getting Started
Quick Start
Get a reliable first dictation setup in about 10 minutes, without random permission or shortcut issues.
Goal
By the end of this page, you should be able to:
- trigger recording reliably
- dictate one sentence into your target app
- understand where to tune behavior next
1. First-Run Setup Checklist
When you open Voiscribd for the first time:
- complete onboarding first - it helps with everything in this setup
- set at least one
Record Shortcut - grant
Microphonepermission - grant
Accessibilitypermission for reliable insert-at-cursor behavior - grant
Screen Recordingpermission if you want screen context and some capture workflows
You can re-check status anytime in Settings -> System Permissions.
2. Set a Trigger That Feels Natural
Go to Settings -> Record Shortcuts, then configure Hotkey 1.
Recommended first trigger:
Right CommandorRight Option- tap: start/stop
- hold: push-to-talk
3. Choose Your First Model
Open Models and set one default model.
Then set language behavior:
- use
Autofor mixed language speech - use fixed language for consistent single-language workflows
4. Run a 30-Second Real Test
- open your real target app (notes/chat/editor)
- trigger recording
- speak one short sentence
- verify output arrives where expected
- stop and repeat once
If this passes twice in a row, your base setup is good.
5. Fix the Most Common First-Day Issues
Recording trigger does nothing
Check System Permissions and hotkey conflicts with other apps.
Text appears in wrong place
Switch to safer insertion behavior first, then return to direct insertion once stable.
Output looks noisy or clipped
Tune Model Studio VAD settings next.
6. Turn On Expert Mode When Ready
Go to Settings -> General -> Expert Mode when you need:
- Spark + Chain controls
- Hot Corners
- deep VAD tuning
- privacy/retention cleanup actions
- diagnostics and local model retention controls
Note
[Screencast Placeholder] Record a full first-launch setup showing each permission prompt, one failed trigger case, and the fix path.