Voxis tour: real-time voice translation
An interactive tour of Voxis: live spoken translation of videos, games and calls on Windows and Linux. 79 languages, 10 free minutes every day. This is the text version of the interactive tour: every chapter’s headline, text and notes.
Subtitles make you read. Voxis speaks.
Real-time · driverless · 79 languages
It translates whatever is playing on your computer into your language, live, and speaks it. Video, games, calls: whatever the language, you hear it in yours, a few seconds behind.
- Not subtitles but a voice: your eyes stay on the screen.
- The delay is a few seconds. That's how simultaneous interpretation works, and we don't hide it.
- The original audio drops (to 30%) while a sentence is spoken, then comes back.
- Basis
- The live site's product copy and the app's own status line: "Translation arrives a few seconds after the speech."
- Limit
- The translation in this scene is dramatized; in the real product the spoken language is detected automatically, and there are 79 target languages.
- Source
- voxislive.com/
You can't read and watch.
Why not subtitles
Subtitles pull your eyes to the bottom of the screen. Voxis speaks: your eyes stay on the scene while you hear it in your language. The source line still sits right above the translation, for anyone who wants to read it.
- The contrast pair: subtitles make you read, Voxis speaks.
- The source line doesn't disappear; it sits above the translation in the app window.
- Basis
- The site's tagline "Subtitles make you read. Voxis speaks." and the two-tier caption (source + translation) in the app.
- Limit
- The "on the scene" percentage is a narrative device, not a measurement; the product does no eye tracking.
- Source
- voxislive.com/
Three steps. A few seconds behind.
How it works
Pick a scenario, pick your language, press Start. You never choose the spoken language; it's detected on its own. The translation trails the speaker by a few seconds. That is the natural delay of simultaneous interpretation, and no setting removes it.
- No source-language picker: Voxis detects it on its own.
- Speech detection runs on your device, translation in the cloud.
- The delay can't be tuned away, and we say so.
- Basis
- The app's welcome tour ("Three steps: scenario, language, Start.") and the download page's requirements: translation runs in the cloud, speech detection on the device.
- Limit
- Connecting usually takes a few seconds, up to half a minute at busy hours.
- Source
- voxislive.com/download
No cable. Nothing to route.
Driverless capture
In Video & Game mode Voxis listens to the system mix directly and leaves its own voice out, so it never translates itself. No virtual audio device to install, nothing to route. Meeting mode is a different story: it needs a virtual microphone.
- Video & Game: zero setup.
- Its own voice is excluded: no echo, and it never translates itself.
- Meeting needs a virtual microphone. We say so here too.
- Basis
- The site's "Driverless capture" section; the download page: "adds no audio device to your system".
- Limit
- On very old Windows 10 builds it falls back to a compatibility mode, and the app says so in the status line.
- Source
- voxislive.com/download
Speak your language. They hear theirs.
Meeting · two way
Meeting mode runs two live translations at once: incoming and outgoing. The other side hears you in their language through a virtual microphone. On Windows that is the free VB-CABLE, and Meeting will not start without it. Teams, Zoom, Meet.
- Two sessions at once: incoming and outgoing.
- The cable is REQUIRED on Windows; without it the app blocks Start.
- No Meeting on the free tier; the one-time 15-minute taste includes it.
- Basis
- The site's Meeting section ("Meeting mode will not start without one") and the app's "Virtual microphone required" card.
- Limit
- The outgoing voice doesn't clone yours; expect the short lag a human interpreter would need.
- Source
- voxislive.com/use-cases/meetings
Nobody joins your call.
No bot · private
Voxis never joins your call as a participant. The only thing that leaves your machine is the audio sent to the translation model. It isn't recorded or shared; only the minutes used are reported, for billing. Telling the other participants is up to you.
- Nothing is added to the participant list.
- Audio leaves for translation only; nothing is recorded.
- In Meeting, informing the participants is up to you. The app reminds you at the start of every Meeting; "Don't show again" turns the reminder off.
- Basis
- The app's Privacy dialog and the site: "No bot ever joins your call… Nothing is recorded or shared."
- Limit
- Audio doesn't stay entirely on your device: it goes to the cloud for translation.
- Source
- voxislive.com/
Pick the language you want to hear. Voxis detects the rest.
79 languages
You choose the language you hear the translation in; the spoken language is detected on its own. The free voice speaks 35 of the 79 languages, and the other 44 arrive as live captions. Paid packs speak all 79.
- You pick the target language, never the source.
- 35 of 79 spoken on the free voice. The rest are captions, not a fault.
- Interface language (23) and translation language (79) are different things.
- Basis
- The site's 35/79 stat strip, the pricing comparison table and the language list the site publishes (voxislive.com/app.json).
- Limit
- The list of free-voice languages can change over time; the current list is on the site.
- Source
- voxislive.com/pricing
15 Pro minutes, once. Then 10 free every day.
Start free
A new account gets 15 minutes of the real human voice, once. After that, 10 free minutes every day in the Windows app and the Linux snap: the same translation in a simpler voice that runs on your PC, Video & Game only. Meeting runs on the minutes in your account. No subscription; minutes you buy never expire.
- The taste is 15 minutes, once; the 10 daily minutes are forever.
- Daily minutes renew at midnight UTC and don't carry over.
- Packs run from Starter, 80 min for $6.99, to Studio, 2600 min for $149.99, VAT included. Students and teachers get 50% off.
- Basis
- The pricing page and the app's "Voxis has three rungs" card.
- Limit
- One daily free allowance per machine; Meeting is not part of the free tier.
- Source
- voxislive.com/pricing
Every session in its own folder.
Transcripts · on-screen text
The translation streams into a live transcript. When the session stops, export it as bilingual TXT, SRT or VTT, each session in its own folder. Watching full screen? The on-screen text sits over the video or game: translation only, while the source line stays in the app window.
- JSON always; TXT, SRT and VTT optional; the folder is Documents\Voxis\Transcripts.
- The on-screen text shows the translation only; 4 looks, Ctrl+Alt+O.
- Audio recording (two separate WAVs) is Video & Game only, and off by default.
- Basis
- The site's "Live transcript & export" and "On-screen subtitles" sections; the app's Settings › Saving.
- Limit
- Subtitle timing is approximate (simultaneous delay); export once the session has ended.
- Source
- voxislive.com/docs/interface-reference
Install. Start. Hear it in yours.
Windows and Linux
Install from the Microsoft Store in one click, sign in, and start with 15 minutes of the Pro voice. On Linux, the same app via snap. No macOS app yet; on a Mac, the Chrome extension works.
- Official sources only: voxislive.com, the Microsoft Store, github.com/VoxisLive/voxislive.
- The Store build brings the system components it needs with it.
- Basis
- The download page.
- Limit
- No Mac app; the waitlist is on the site.
- Source
- voxislive.com/download
Next steps
Get it on Microsoft Store Pricing Home
Other pages: Intro · Setup · App tour
Language: English · Türkçe · Español · Deutsch · Français · Português · 한국어