Caption styling is available from Starter upwards, in Branding. Free clips are captioned too, in the default style. Every change applies to your next talk, and the panel shows a live preview using a frame from your last one.
The fifteen styles
Eight of these arrived in September 2026. They differ from each other in more than typeface: one moves a shape along the line, one types, one stacks, one glows, one inverts the whole thing to dark words on a light card.
| Style | What it does |
|---|---|
| Anthem | Two-word statements that land large. Consecutive lines alternate white and your highlight colour, so a run of them reads as a rhythm. |
| Rider | The line holds still and a rounded pill rides along it, word by word. The word under the pill turns dark as the pill arrives. |
| Typewriter | Characters type onto a bar behind a cursor. The bar is sized to what is written on it, so it opens under the first word and widens as the line fills. |
| Ladder | Words stack down the left edge, a sentence at a time, and the one the sentence turns on is set large. Small joining words tuck onto the line below. |
| Neon | The words are a lit tube: a coloured glow around a hot core. |
| Hollow | The line stands as empty outlines and each word fills in solid as it is spoken. |
| Sticker | Dark words on a light card, set slightly askew with a tape tab. The only style that inverts light and dark, which is what makes it read differently beside the others. |
| Drift | Small tracked capitals that arrive scattered and settle. The quietest of the new set, and the most cinematic. |
| Editorial | Compact, closely spaced text with its own text colour, shadow, line height and letter spacing controls. |
| Karaoke | The old default. A few words sit on screen and each lights up as it is spoken. |
| Word Pop | One word at a time, in your highlight colour, held for exactly as long as it is said. |
| Highlight Box | One word at a time, sitting on a filled box. |
| Boxed | Karaoke on a solid bar behind the text. The most legible over busy footage. |
| Build Up | Words appear one by one and stay, so the sentence assembles itself. |
| Clean | Plain text, nothing highlighted. The quietest option. |
Word Pop and Highlight Box show one word at a time by their nature, so the words-per-view setting does not apply to them.
Break every caption across two lines is a setting rather than a style, so it applies to whichever style you have picked. Three styles cannot take it, because they already put one word on each line: Word Pop, Highlight Box and Ladder. The page says so when you are on one of them.
Settings a style brings with it
Common controls include font, size, colours and placement. Some layouts limit words per view or two-line wrapping, as noted above. Six styles add controls of their own, which appear under the picker when you choose them and disappear when you choose something else. Your settings for a style are kept while you try another one.
| Style | Its own settings |
|---|---|
| Rider | The pill's roundness, width and height. |
| Typewriter | The bar's side padding, height, colour and opacity. |
| Ladder | Line height and letter spacing, and which word is set large: the strongest one on screen, only words you list, or never. |
| Editorial | Text colour, shadow colour and strength, line height and letter spacing. |
| Neon | The glow's colour and opacity. |
| Sticker | The card's angle, colour, width and height. |
Ladder's automatic choice is worth understanding, because it is the one that can look wrong. It scores the words on screen together and sets the strongest of them large, so a better word is never passed over for a weaker one sitting beside it. When nothing on screen is worth pointing at, it points at nothing rather than picking the least bad option. On a typical talk it accents about one screenful in three.
Colour
Two colours do most of the work:
- Highlight, the word being spoken. Yellow by default, because it survives almost any background.
- Inactive, everything else. White by default.
Boxed and Highlight Box add a third, the box colour, which only appears as an option when one of those two styles is selected.
Three of the newer styles take their text from Inactive in a way worth knowing, because there is no separate word being spoken to colour: Neon lights its letters in it, Hollow draws its outlines in it, and Sticker sets its words in it. On Sticker there is one exception. Its card is light and Inactive is white by default, so a white-on-white caption nobody could read would be the out-of-the-box result. When your text colour would disappear into the card, we keep it dark instead. Any colour you actually choose is used as you chose it.
If you are matching brand colours, check the result over your actual footage in the preview rather than against white. A colour that sings in a brand guide can vanish against a dim stage.
Font, size and casing
Twelve fonts are built in: Integral (the default), Anton, Bebas Neue, Archivo Black, Montserrat, Poppins, Oswald, Sora, Inter, Heros, DM Serif Display and Playfair Display. On Growth you can upload your own .ttf or .otf, and it then appears as Custom.
Size is a slider, running from 3.5% to 12% of the frame's short side. Because it is a percentage rather than a pixel count, the same setting looks the same on every clip. The default is 4.2% and suits most talks, and the desktop app carries the same slider. The old names (Small, Medium, Large and Extra large) still apply to clips and saved settings made before the change, where they map to the percentage they always meant.
Casing can be UPPERCASE (the default), lowercase, Title Case, or As spoken, which keeps whatever the transcript produced.
Words on screen at once
Between one and six, four by default. Fewer words means bigger, punchier text and more cuts; more words means a calmer read and fewer changes. Four is a good middle for preaching, where sentences run long.
Where they sit
A slider moves the caption band up or down the frame, anywhere from 20% to 85% of the way down.
The range deliberately stops short of both edges. Too high and captions collide with a top-corner logo; too low and the platforms draw their own buttons, usernames and progress bars straight over your words. If you are unsure, leave it near the middle.
Words we say Growth and up
Every church has words no transcription gets right first time: your pastor's name, your building, a Yoruba or Igbo phrase, the name of your Tuesday night gathering. Words we say, in Branding, is a list of those words, and it is applied to every new clip.
- Add words one at a time, or paste a comma-separated or line-separated block to add many at once.
- For a word that comes out consistently wrong, you can record what it is misheard as, which makes the correction more reliable.
- Corrections you make by hand on a single clip are left alone.
It takes ten minutes once and then quietly saves you a correction every week. Filling this in is the highest-value thing a new Growth account can do.
What we keep fixed
Each style defines its own layout, animation and readability treatment. Controls vary by style: Ladder and Editorial expose line height and letter spacing, and Editorial also exposes shadow controls. Settings not shown for the selected style use its built-in design.
These were tuned against real sermon footage on real phones. If you have hit a case where one of them genuinely does not work, tell us: that is a better bug report than it sounds.