How to Set Up a Floating Live Caption Overlay on Mac

The short answer

To read live captions over a video, call or other Mac app, use a separate caption window and arrange it around the content you need to see. For the spoken language alone, first check the app's own captions or Apple's Live Captions. For original speech and a supported translation together, a system-audio app such as Alison Workspace can display a floating bilingual overlay.

The useful setup has three parts: choose the audio source, prepare recognition and translation, and make the window readable. A floating window does not add a subtitle track to the video or send translated speech to another person. Its behavior in full screen, different desktop Spaces and screen sharing needs a check with the apps you actually use.

Choose captions, translation or a subtitle file

A video's supplied subtitle track is usually the first choice when it is available and accurate. It follows the creator's timing and may include information that speech recognition misses. Live captions instead create text from speech while audio plays. Translation is another step, with its own language requirements and possible errors.

Apple's Live Captions provides a built-in floating reading window on supported Macs, subject to language and regional availability. If you need a translation, compare the player or meeting's language options before adding another app. Alison Workspace is a macOS app that recognizes accessible playback audio or a selected microphone and can show original and translated text in its own window.

Start with one predictable audio source

Use a short, clear clip before arranging the whole workspace. Alison Workspace requires macOS 15 or later and Apple silicon, M1 or newer.

1. Open Alison Workspace and finish its initial setup.
2. Choose system audio for a video, podcast or remote voice playing on the Mac. Choose microphone for nearby speech; use both only when you need both sources.
3. Allow the permission requested for the selected input. System audio and microphone access are separate permissions.
4. Select the recognition language to match the speaker, and choose a different supported translation target if you need translated text.
5. Prepare any required language resources, start the clip, and choose Start Listening.
6. Make sure subtitle display is enabled in the menu-bar controls.

If no text appears, solve the input or recognition problem before changing the window layout. Mute other clips and notification sounds so the example is easy to judge.

Place and resize the floating window

Move the window using its upper control rail, then resize it to leave the player controls, slides and speaker visible. A wide, shallow window can work for a video; a taller window may help when you need more recent text beside notes. Choose a position that reduces eye movement rather than covering the center of the content.

In subtitle settings, adjust text size and style to suit the actual screen and viewing distance. If a sentence wraps too often, try a wider window before making the type smaller. Check names, numbers and punctuation at the size you intend to use. A layout that looks attractive while paused may be difficult to read during rapid speech.

Keep original and translated text easy to compare

Select bilingual display when you want both lines. Alison's subtitle settings also provide original-only and translation-only modes, plus the order of the two languages. For technical content or language learning, keeping the original prominent makes a doubtful translation easier to check.

The original and translation may arrive at different moments as recognition settles and translation follows. A momentary delay is different from a missing translated line throughout the session. Avoid judging the translator from a half-finished word. Test a complete sentence containing one name and one number, and replay it if either line is uncertain.

Understand click-through interaction

A caption window should not make the underlying app difficult to use. Alison has a click-through setting. Its current overlay becomes interactive when the pointer is over the window, so its controls and movement area can still be used. Move the pointer away after adjusting it and test the underlying player's buttons.

If a click does not reach what you expect, move the overlay clear of the control first. Check the click-through setting rather than repeatedly clicking through important meeting controls. Click-through concerns mouse input; it does not make the window invisible to a screen-sharing tool.

Check full screen, Spaces and a second display

macOS gives full-screen apps their own desktop Space. A caption window that is visible beside an ordinary browser window therefore needs another check after you enter full screen or change Spaces. Alison's overlay is configured as a floating panel intended to remain useful across these arrangements, but that is not a guarantee for every application or display setup.

Test windowed playback first, then enter full screen, switch to another app and return. If the overlay becomes hard to find, leave full screen, check subtitle display, and move it back beside the content. With an external display, place the player and captions deliberately and repeat the check after reconnecting the display. Use windowed playback if that is the reliable layout on your Mac.

Separate personal reading from screen sharing

The text can be processed locally and still appear in a shared screen. Before presenting, decide whether the audience should see the captions. Sharing the entire display is a reason to treat every visible caption as potentially visible to others. Sharing a selected window offers a more limited sharing scope, but the receiving view is what matters.

Run a short rehearsal with a permitted test sentence and inspect the view from a second device or participant. If you cannot verify it, hide the overlay before sharing sensitive content. Tell participants when captions or translation are in use and follow the applicable meeting rules.

Choose a layout you can use for the whole session

Finish with a one-minute trial in the actual browser, player or meeting app, using the usual headphones and display arrangement. Confirm the audio source, both language lines, readable type and reachable controls. Automatic text is a comprehension aid: overlapping speakers, music, unfamiliar names and protected playback can still cause problems.

Alison's documented workflow processes audio locally after the necessary resources are prepared, does not retain raw live audio, and keeps local text history until you delete it. For a floating translator, test the audio route and language pair as carefully as the window. Use the related system-audio and bilingual-caption guides to choose the right workflow before relying on it.