Frequently Asked Questions

VoxInk is a voice input tool for Mac and Chrome. It has four modes: Direct (voice to text), Polish (voice to clean writing), Translate (speak one language, output another), and Listen (real-time translated captions for calls and meetings).

macOS Dictation gives you raw transcription with a 30-60 second timeout. VoxInk adds AI polish (clean up filler words and grammar), real-time translation to other languages, and Listen mode for understanding foreign language conversations. No timeout — talk as long as you want.

VoxInk recognizes 60+ spoken languages automatically. Translation output is available in 17 languages including English, Chinese, Thai, Vietnamese, Indonesian, Malay, Filipino, Japanese, Korean, Arabic, and more. See full list →

No. VoxInk requires an internet connection for speech recognition and translation. Audio is streamed in real-time and never stored.

Yes. Listen Mode captures all audio playing through your Mac — any app that plays sound will be captured. This includes Zoom, Google Meet, Teams, WeChat, LINE, FaceTime, YouTube, and more.

No. VoxInk captures audio at the system level without joining the meeting as a bot or participant. It's completely invisible to others. No notification is sent.

Yes. When you start Listen Mode, a web link is generated. Share it with participants — each person can open it and choose their preferred language. Everyone sees the conversation translated to their language from one VoxInk session.

Recording laws vary by jurisdiction. Many places require all-party consent. We recommend informing participants that the conversation is being transcribed. VoxInk provides transcript text, not audio recordings. Check your local laws. Enable Privacy Mode if you don't want any record saved.

No. Audio is never stored — not on your device, not on our servers. Audio is streamed in real-time to the transcription engine and discarded immediately after processing. This is true regardless of whether Privacy Mode is on or off.

When Privacy Mode is on, the resulting transcript text is also discarded — nothing is saved locally or in the cloud. When Privacy Mode is off, transcripts are saved locally on your Mac for later reference. Learn more →

Speech-to-text: Soniox. AI polish and translation: Google Gemini. Backend proxy: Cloudflare Workers. Neither service retains your audio data after processing.

VoxInk has a free tier: 60 minutes of Direct mode, 10 minutes of Listen/Polish, and 5 minutes of Translate. After that, upgrade to Plus ($2.99/month) or Pro ($6.99/month) for more minutes. See pricing →

Yes. One account, one subscription, both platforms. Sign in with the same Google account on both and your quota is shared.

Yes, but with limited functionality. Without signing in, usage can't be tracked across devices and quota management is limited. We recommend signing in for the best experience.

Check: (1) Microphone permission is granted in System Settings → Privacy → Microphone. (2) The correct microphone is selected in VoxInk Settings. (3) If using noise cancellation, try turning it off — your mic may be too far. (4) Make sure VoxInk is running (green icon in menu bar).

The first time you use Listen Mode, macOS will ask for "Screen & System Audio Recording" permission. Grant it in System Settings → Privacy → Screen & System Audio Recording → VoxInk. You may need to restart VoxInk after granting.

VoxInk requires Accessibility permission to detect hotkeys globally. Go to System Settings → Privacy → Accessibility and make sure VoxInk is enabled. If it's already enabled, try toggling it off and on.