Subtitles & Display Options
Customize the karaoke subtitle overlay and how the transcript is displayed.
Kalima's Studio lets you tailor how a transcript looks and how captions are presented, from a synced karaoke subtitle overlay to text size, layout, and a set of ready-made caption styles. This page shows you how to set everything up so your transcript reads clearly on screen, on a projector, or in a small floating window.
- Karaoke subtitle overlay: Captions that track the audio word-by-word, in fullscreen or a floating window.
- 7 style presets: Pick a caption look, from clean and minimal to high-contrast and cinema.
- Flexible layout: Read original and translation side-by-side or stacked, and filter what's shown.
- Readable text: Resize the transcript from 12px to 64px and go fullscreen for presenting.
The karaoke subtitle overlay
When you play back a recording, Kalima can show the spoken text as a clean, synced subtitle. As the audio plays, the word being spoken is highlighted with a green box, words already spoken stay bright, and upcoming words are dimmed, exactly like karaoke. This makes it effortless to follow along, present captions to a room, or keep a caption visible while you do something else.
The overlay comes in two variants:
- Fullscreen, larger caption text for presenting to an audience or focused review.
- Picture-in-picture, a smaller caption that stays visible in a floating mini-window while you work elsewhere.
The karaoke subtitle shows only readable speech, control markers in the transcript are filtered out so the caption stays clean. When playback is paused, every word shows at full brightness instead of dimming the upcoming text.
Following along while you listen
The same word-by-word highlighting also runs inside the regular transcript while audio plays. You can click any word in the transcript to jump playback straight to that moment, and the highlight will continue from there. Both features rely on word-level timing captured during the recording.
Subtitle style presets
Choose from seven caption styles to match your video, presentation, or accessibility needs. Each one shows a small preview swatch in the style selector so you can see its colors before applying it. Your choice is remembered for next time.
| Preset | Look | Best for |
|---|---|---|
| Classic | Default caption card | General use |
| Translucent | Semi-transparent overlay | Captions over busy backgrounds |
| Cinema | Dark background, gold text, movie look | Captions over video |
| Minimal | Borderless, clean text | Distraction-free reading |
| Outline | Text with a shadow outline | Captions over varied imagery |
| High-contrast | Bold, accessible black and white | Accessibility and visibility |
| Frosted | Blur and glass effect | A modern, layered look |
For accessibility, or any time the caption sits over moving or unpredictable video, High-contrast and Outline keep text readable no matter what's behind it.
Transcript display options
Beyond captions, the transcript toolbar gives you several ways to control how text appears while you read, edit, or present.
Layout: split vs. stacked
When a session has a translation, you can toggle between a split view (original and translation in two side-by-side columns) and a stacked view (translation shown beneath the original). The layout toggle is the leftmost button in the transcript toolbar, and each click switches between the two.
Split view is most meaningful when a translation exists; with no translation, there is simply one track to read. Learn more about adding a translation in Live Translation.
Display filter
A three-button segmented control, All, Original, Translation, lets you show everything, only the transcribed source text, or only the translated text. Use it to focus on just the language you need without the other one distracting you. The Translation option is only useful when the session has translations.
Font size
Click the text-size (A) button in the toolbar to open a slider, then drag it to set the transcript text anywhere from 12px to 64px. Larger text is great for presenting or low vision; smaller text shows more at once.
Fullscreen
The fullscreen button expands the transcript to fill the whole screen for distraction-free reading or presenting a live transcript to a room. Press the fullscreen button again, or press ESC, to exit.
Language badges
Small language-code badges (such as en, es, or fr) appear in segment headers to show the detected language, and segments that mix languages highlight which spans belong to which language. This makes multilingual and code-switching conversations easy to scan. See Languages & Detection for how detection works.
Your display preferences, subtitle style and font size, are saved in your browser, not your account. They won't follow you to another device or browser, so you may need to set them again there. This is expected behavior.
Frequently asked questions
Seven: Classic, Translucent, Cinema, Minimal, Outline, High-contrast, and Frosted. Open the subtitle style selector and pick one, a preview swatch shows its colors, and your choice is remembered for next time.
Yes. Open the text-size control in the transcript toolbar and drag the slider. You can go from 12px up to 64px, and your choice is remembered in that browser.
These preferences are saved in your browser rather than your account, so they stay on the device and browser where you set them. Just set them again on the other device, this is expected.
Word-by-word karaoke highlighting and word-click seeking both rely on word-level timing from the recording. Make sure the recorded audio is loaded and playing; if timing isn't available, highlighting and click-to-jump won't run.
When a session has a translation, use the layout toggle (leftmost button in the transcript toolbar) to switch between a side-by-side split view and a stacked view. The display filter also lets you show only the original or only the translation.
They show the language Kalima detected for each segment (for example en or es). In segments where languages mix, the relevant spans are highlighted so you can tell which words belong to which language.