What the overlay is
The overlay is an optional always-on-top window that shows the translated line while a session runs. It floats above the app you are actually using — the video player, the browser tab, the conferencing window, the game — so you can read the translation without giving up the thing you are reading it for. The main VoxisLive window can stay behind it, minimised, or on a second screen.
It is a reading surface, not a second control panel. It carries the translated line and nothing else: no buttons, no settings, no menu of its own. Everything that decides what appears in it — the target language, the scenario, the voice — is set once in the main window before you start, and the overlay simply reflects the session that is already running. If you have not read how it works yet, that page covers the capture and translation path the overlay is displaying the end of.
One keyboard shortcut opens and closes it
The overlay is bound to a keyboard shortcut: one press opens it, another press closes it. That matters more than it sounds like it should, because the moments when you want captions are the moments when you cannot go hunting for a window — mid-call, mid-scene, mid-match.
The shortcut works while another application has focus, which is the point — you do not have to click back into VoxisLive first. You can leave the app running behind a full-screen video, hit the key when a line goes past you, and hit it again when you are done.
This is the one piece of overlay behaviour that appears on no other page today: the overlay has a keyboard shortcut and it toggles — open and close from the same key. If you have used VoxisLive and never found the overlay, that is most likely why.
It shows the target language only
The overlay is target-language focused. It shows the translation — the line in the language you chose to hear — and it does not show the original. That is a deliberate split, not an omission: the overlay is meant to be glanceable, and putting two languages in a floating window makes it something you study rather than something you glance at.
The source line is not lost. It lives in the main window, where it sits above the translation bubble, so the pairing is there when you want it. The overlay is what you read while you are doing something else; the main window is where you look when you want to compare the two languages or check what was actually said.
The same split carries into what you keep afterwards: the saved transcript holds both languages, and exports can be bilingual or translation-only, so the overlay's single-language view never decides what ends up in the file. See the features overview for where transcripts and exports sit.
In games
The overlay is used a lot over games, and the reason it works there is the same reason the rest of VoxisLive works there: we operate on the audio path, not inside the game. VoxisLive reads what the system is already playing and renders the translation separately, so the game's own sound is unaffected.
Drawing a window on top of a game is a different problem from capturing its audio, and this is where the honest limitation lives. In full-screen modes, drawing the overlay window is not guaranteed. Exclusive full-screen in particular hands the display over to the game, and an ordinary always-on-top window may simply not be composited over it. The translated audio keeps playing either way — that path does not depend on the overlay at all — but the text may not appear.
- Borderless windowed or windowed mode is the configuration where an always-on-top window behaves predictably.
- Exclusive full-screen may hide the overlay entirely; this is a property of how the display is handed to the game, not a setting inside VoxisLive.
- A second monitor sidesteps the question — the overlay is a normal window and can sit on the other screen.
- If the overlay does not draw, the translated voice and the saved transcript are unaffected.
S1/S2 labels when the speaker changes
When more than one voice is present, the overlay can mark the change with a short S1 / S2 label. This is speaker-change detection: it notices that the voice has changed and labels the lines accordingly. It does not know who anyone is, it does not match a voice to a name, and it does not carry an identity between sessions. S1 simply means “a different voice from S2”.
The labels appear when a change is detected. A single-speaker stream — one narrator, one presenter, one person talking to camera — stays unlabelled, which is correct behaviour rather than a failure. On two-way meeting calls, where the two directions are already separated, the labels add one more layer of structure.
A file output for OBS
For streaming there is a file output for OBS. Rather than asking a broadcast tool to capture a floating window — which is fragile, resizes badly and puts your own reading surface on your viewers' screens — VoxisLive writes the caption line to a file that OBS can read directly with its text source.
That separation is the point. A file source is a first-class OBS input: you style it with OBS's own font, colour, outline and positioning controls, place it where your layout wants it, and show it in one scene and not another. Your overlay window and your viewers' captions stay independent — the size you read comfortably is rarely the size a viewer needs on a phone.
- Start a VoxisLive session as usual and enable the file output.
- In OBS, add a Text source and point it at the file VoxisLive is writing.
- Style and position the text source in OBS — the appearance of the stream caption is decided there, not in VoxisLive.
- Keep the overlay window off-stream, on your own screen or a second monitor, if you also want to read along.
What the overlay does not do today
The overlay is deliberately small, and the parts that are missing are worth naming plainly instead of leaving you to look for settings that are not there. None of the following exist today.
| Not available today | What that means in practice |
|---|---|
| Font size | There is no control for caption text size in the overlay window. The text renders at its built-in size. |
| Transparency | There is no opacity slider — you cannot make the window semi-transparent so that what is behind it shows through. |
| Display mode | There is no choice of presentation style — no banner-versus-box mode, no compact or expanded variants. |
| Position memory | The overlay does not remember where you last put it, or which monitor it was on, between sessions. |
If you have used a subtitle tool with a panel full of typography controls, this reads as a short list of gaps, and that is a fair reading. The overlay today is a window that shows the translated line and gets out of the way. When those settings arrive they will show up in the changelog; until then, this page will keep saying they are absent.
Two of these have a workaround worth knowing. If you need caption text you can style, use the OBS file output — OBS's own text source gives you font, size, colour and placement. And if you need the overlay on a particular part of the screen every time, a second monitor makes the lack of position memory much less of a daily annoyance.
Where the overlay fits
Most sessions are primarily audio: you hear the translated voice a few seconds behind the original, and that is the product. The overlay is the reading channel alongside it, and it earns its place in three situations — a room too noisy to trust your ears, material dense with names and terms you would rather see written, and a stream where captions are for an audience rather than for you.
It costs nothing extra: the overlay is part of the app rather than a separate tier, and it behaves the same on every plan. See pricing for what a session costs, or download to try it against something you are already watching.