Getting Started with EdgeWhisper: A Complete Guide
Learn how to install, configure, and master EdgeWhisper for macOS. From your first dictation to advanced features like personal dictionaries and snippets.
EdgeWhisper Team
Product Documentation
Getting Started with EdgeWhisper: A Complete Guide
Welcome to EdgeWhisper — the fastest way to turn your voice into text on macOS. This guide walks you through everything from installation to advanced productivity techniques.
Installation
Requirements
- macOS 14 (Sonoma) or later
- Apple Silicon Mac (M1, M2, M3, or M4)
- ~2 GB free storage for the Voxtral AI model
Install from the Mac App Store
- Open the Mac App Store
- Search for "EdgeWhisper"
- Click Get → Install
- Open EdgeWhisper from your Applications folder
On first launch, EdgeWhisper will download the Voxtral Mini 4B model (~1.5 GB). This only happens once — after that, everything runs offline.
Granting Permissions
EdgeWhisper needs two system permissions to function:
Microphone Access
When prompted, click Allow to grant microphone access. EdgeWhisper processes all audio locally — nothing is sent to any server.
Accessibility Access
EdgeWhisper uses macOS Accessibility to type text into any application. Go to:
System Settings → Privacy & Security → Accessibility → Enable EdgeWhisper
This is the same permission used by text expanders and keyboard utilities. It allows EdgeWhisper to inject the transcribed text wherever your cursor is.
Your First Dictation
- Place your cursor in any text field (email, document, code editor, etc.)
- Press your hotkey (default:
⌥ Option + D) to start dictation - Speak naturally — EdgeWhisper handles punctuation, filler words, and formatting
- Press the hotkey again or pause to stop dictation
Your text appears in real-time as you speak. EdgeWhisper's AI automatically:
- Adds punctuation (periods, commas, question marks)
- Removes filler words ("um", "uh", "like")
- Capitalizes sentences and proper nouns
- Formats numbers and common abbreviations
Tips for Better Dictation
Speak Naturally
Don't try to speak in a special "dictation voice." Voxtral is trained on natural speech patterns. Speak at your normal pace and cadence — the AI handles the rest.
Use Verbal Punctuation
While EdgeWhisper auto-punctuates, you can be explicit when needed:
- Say "period" for
. - Say "comma" for
, - Say "question mark" for
? - Say "new line" for a line break
- Say "new paragraph" for a paragraph break
Optimal Microphone Setup
- Use your Mac's built-in microphone or an external USB microphone
- Keep the mic 6-12 inches from your mouth
- Minimize background noise when possible (though Voxtral handles moderate noise well)
- AirPods and Bluetooth headsets work, but wired connections offer lower latency
Advanced Features
Personal Dictionary
Teach EdgeWhisper your specialized vocabulary:
- Open EdgeWhisper preferences
- Go to Dictionary
- Add custom words, names, or technical terms
- EdgeWhisper will prioritize these in transcription
This is essential for medical terminology, legal jargon, brand names, or any domain-specific vocabulary.
Snippet Library
Create text shortcuts that expand when spoken:
| Trigger | Expansion |
|---|---|
| "my email" | raphael@elitizon.com |
| "company address" | 123 Innovation Street, London, UK |
| "sign off" | Best regards,\nRaphaël MANSUY |
Snippets save you from repeatedly dictating common phrases.
Supported Languages
EdgeWhisper supports 13 languages via Voxtral:
- English
- French
- German
- Spanish
- Italian
- Portuguese
- Dutch
- Hindi
- Arabic
- Chinese
- Japanese
- Korean
- Russian
EdgeWhisper auto-detects your language — no configuration needed. You can even switch languages mid-sentence.
Keyboard Shortcuts
| Shortcut | Action |
|---|---|
⌥D | Start/Stop dictation |
⌥⇧D | Dictation with auto-edit |
⌘, | Open preferences |
Esc | Cancel current dictation |
Troubleshooting
"No microphone access"
Go to System Settings → Privacy & Security → Microphone and ensure EdgeWhisper is enabled.
"Model not downloaded"
Check your internet connection and restart EdgeWhisper. The initial model download requires ~1.5 GB.
Text not appearing in certain apps
Some sandboxed applications may block Accessibility input. Try granting the app explicit Accessibility permissions.
High CPU usage on first use
The Voxtral model takes a few seconds to load into memory on first dictation. Subsequent dictations are instant.
What's Next?
- Explore the AI auto-edit features in EdgeWhisper Pro
- Build a personal dictionary for your domain
- Set up snippet libraries for common phrases
- Join our community for tips and updates
Have questions? Reach out at contact@elitizon.com — we'd love to hear from you.