| name | Trending caption styles |
| description | Apply modern short-form caption looks — word-pop, karaoke highlight, bold-boxed, typewriter — sized and placed for 9:16. |
Trending caption styles
Follow the build-in-kaestral house rule: add captions as real text clips ON the Kaestral timeline via add_texts/update_text (the user watches them appear); never burn captions into a standalone file outside Kaestral. The final video comes from export_project(mode:"video").
Captions drive retention (most short-form is watched muted). Kaestral renders and animates text via add_texts/update_text with textStyle + a textAnimation preset. Below are the looks that read as "modern short-form" and exactly how to build each.
Placement + sizing (do this for all styles)
- Vertical (9:16): place captions in the LOWER-MIDDLE, above the platform UI safe zone (roughly 12–20% up from the bottom). Hooks go TOP third.
- 2–5 words per screen. Never a full paragraph. Break long lines.
- Size at 1080p: captions ~60–80px, hooks ~90–120px. Bold. High legibility.
- Always ensure contrast: a background box OR a thick outline/shadow so text survives any footage.
Styles (pick to match the vibe)
1. Word-pop (energetic, most common)
textAnimation preset wordPop (or wordReveal for a smoother reveal).
textStyle: bold, white fill, thin dark outline (border) + soft shadow. No box, or a subtle one.
- Each word pops in on its beat — pair with
beat-sync-cutting for perfect timing.
2. Karaoke highlight (the word being said lights up)
textAnimation preset highlightPop or highlightBlock.
highlightPop: the active word scales + tints to an accent color. highlightBlock: a colored box sweeps behind the active word.
- Set the accent from the palette (
extract_palette) or brand. Great for "read-along" energy.
3. Bold-boxed (clean, punchy, brand-safe)
textStyle with background.enabled = a solid box (brand or palette color), white bold text, popIn animation.
- Reads well on busy footage; strong for hooks and CTAs.
4. Typewriter (suspense / storytelling)
textAnimation preset typewriter. Text types out char-by-char.
- Use for a slow reveal, a quote, or a build-up line. Pair with a calmer section.
5. Slide-up lower-third (informational)
textAnimation slideUp (or wordSlide), left-aligned, smaller, with a subtle bar.
- For names/titles/context rather than spoken captions.
Timing captions to speech (automatic)
add_captions transcribes the clip on-device (whisper) and lays word-accurate caption clips automatically. Give it the spoken clip's clipId, a wordsPerCaption (2–4 for karaoke feel), and the textStyle + textAnimation you want (e.g. highlightPop for karaoke, wordPop for energy). Each caption lands at the exact frames the words are spoken.
- Prefer this over hand-placing text for spoken captions — it's word-accurate. For NON-spoken call-outs (a hook line, a CTA), use
add_texts directly.
- To see the words first (to pick a hook line or edit wording), call
get_transcript.
Apply
For spoken captions: add_captions(clipId, wordsPerCaption, textStyle, textAnimation). For manual call-outs: add_texts (lands on a dedicated text track), then update_text to refine. Keep ONE caption style consistent across the video.