Nine caption families across 45 presets
Versely's 45 caption presets are nine families of five. Map them by backing, position and word timing to shortlist three candidates in a couple of minutes.
Forty-five caption presets sounds like a browsing problem. It is not, because the 45 are nine families of five, and the five inside a family differ only in colour, position or corner radius. Pick the family and you have made the decision that matters; pick the variant and you have made a two-minute one.
Everything below is a map from content type to family, so you can shortlist three candidates without scrolling the full caption styles list end to end.
The nine families
| Family | Font | What the five variants change | Built for |
|---|---|---|---|
| Classic | TikTok Sans | text colour, stroke colour, top or bottom | Talking heads, voiceover, general social |
| Paper | Montserrat | card colour, corner radius, top or bottom | Tutorials, listicles, text-heavy captions |
| Ink | Inter | card colour, corner radius, top or bottom | Cinematic vlogs, aesthetic b-roll, travel |
| Halo | Poppins Bold | text colour, ring colour, top or bottom | Gym content, street interviews, outdoor |
| Hype | Oswald | stroke colour only | Gym and sports, promo, comedy skits |
| Signature | Pacifico | text colour, stroke colour, bottom or centred | Storytime, personal vlogs, lifestyle |
| Headline | Bebas | text colour, stroke colour, three positions | Hooks and openers, news-style recaps |
| Snap | Anton | accent colour only, word-timed | High-retention hooks, ad scripts |
| Reveal | TikTok Sans | accent colour only, word-timed | Storytelling reveals, suspenseful voiceover |
Seven of the nine are subtitle families: the whole line appears and updates together, timed to the voiceover. Two — Snap and Reveal — are word-timed, meaning each word animates in sync with the audio rather than the line arriving as a block.
Each family has a base variant that carries its name, and it is the one to judge the family by. Every family also has its own group page collecting the five variants side by side — the Hype group is a good example of how little separates the members of a family once you are inside one.
Three axes that pick the family for you
Rather than reading nine descriptions, answer three questions.
1. How hostile is the footage?
This is the first cut and it removes most of the list.
- Chaotic, bright, fast-moving — crowds, outdoor shoots, gym floors, hard cuts. You need a heavy ring around the letters. Halo is built for exactly this; Hype is the loud version of the same idea.
- Busy but not hostile — normal talking-head footage with a real background. A thin stroke is enough. Classic.
- Anything, including footage you have not seen yet — a solid card guarantees contrast regardless of what is behind it. Paper on light-friendly work, Ink when a white card would break the mood.
2. Are the words content, or are they a transcript?
- If a viewer can follow the video with the sound on and the captions are there for the mute case, the captions are a transcript. Choose for legibility and get out of the way: Classic, Paper, Ink.
- If the words are carrying the hook — if the caption is the thing being watched in the first two seconds — choose for performance: Snap, Reveal, Headline, Hype.
3. What is the voice of the channel?
- Broadcast or editorial → Headline, which reads as a title card or chyron rather than a subtitle.
- Personal and handwritten → Signature, centred cursive, the one that reads like a note rather than a caption.
- Loud and competitive → Hype.
- Neutral, so the footage carries the identity → Classic or Ink.
Three answers usually converge on two or three families. That is your shortlist, and it is where the browsing should start rather than where it ends.
Word-timed is a different decision from styled
Snap and Reveal are worth separating out, because they are not "Classic with animation". They change the pacing of reading, which changes how the video is watched.
Snap scales each spoken word up and recolours it as it is said. It is the punchiest option in the set and it is built to match fast delivery. Against a slow, considered read it looks frantic.
Reveal builds the line one word at a time and leaves the words on screen. It mirrors a voiceover walking toward a punchline, which is why it suits "wait for it" structures and suspense. Against a fast read it lags behind the audio and feels sluggish.
So the choice between them is not aesthetic, it is a question about your delivery speed. Fast talker, Snap. Building to something, Reveal.
Both families vary on one field only: the accent colour, which is the colour a word jumps to as it is spoken. The size of the effect is literally the distance between the resting text colour and that accent. The Mono variants in each family set both ends to the same colour, so the word still scales or appears but never changes hue — the quietest possible version of a word-timed caption, and the one to reach for when you want the timing without the flash.
The variant is where the brand lives
Once the family is fixed, the five variants exist to let a family fit a brand without leaving it. Read the table above column three and you will see how little actually changes: Hype varies on stroke colour alone, Snap and Reveal on accent colour alone, and the card families on card colour, corner radius and position.
That is a feature. A house style is a family plus one variant, held across every video, and switching families between posts throws away the recognition that captions do in a feed before anyone reads a word. Keeping captions on brand across every video is the operational version of that.
Two things to check when you pick the variant:
- Contrast is measured against the caption's own backing, not against your footage. A card family's ratio is text against its card; a stroke family's is text against its stroke. The translucent "Sheer" variants score lower than the solid card they derive from, which is the trade they make on purpose. None of that is a promise about legibility over footage nobody has seen — that is still your job.
- Position moves more than you think. Several families offer a top variant, and moving captions above the subject is the difference between covering a face and framing it. On vertical video it also keeps the words clear of the platform's own interface furniture at the bottom of the frame.
From shortlist to locked style
- Answer the three questions and take the two or three families they converge on.
- Look at the base variant of each, not the whole group. The base is the family's argument.
- Preview the finalists on five seconds of your own footage rather than on a demo clip. A style that reads well on stock behaves differently against your lighting, your framing and your speaking pace.
- Pick the variant last, and pick it for your brand colours rather than for contrast alone.
- Write it down. Family, variant, position. Three words in a brief is the whole spec.
If the font is the sticking point rather than the family, the nine families draw on eight typefaces — Classic and Reveal both set in TikTok Sans — and each of the eight has its own page. Anton behind Snap and Bebas behind Headline are the two that most change the character of a caption, and the font registry has the rest with previews.
FAQ
Are all 45 presets available on every video?
The 45 are Caption Studio's catalog, and they attach to auto-timed captions generated from the audio. Note that Versely's subtitle tool carries a separate preset list of its own, with its own names and its own tiers. The two are different lists — they share exactly one name, handwritten, which means a different look on each side — so name a style in the vocabulary of the surface you are actually working in.
Which families work best when I do not know what the footage looks like yet?
The card families, Paper and Ink. A solid card behind the text guarantees a fixed contrast ratio no matter what is behind it, which is why they are the safe default for batch work, client footage, or anything you are styling before you have seen every clip.
Can I use a word-timed family on a video with no voiceover?
There is nothing to time to. Snap and Reveal animate against spoken words, so on a silent clip or a music-only edit they have no signal to follow. Use a subtitle family with static text, or a fixed overlay if the words are yours rather than transcribed.
Is switching variant within a family a different kind of job?
No. A variant differs from its family's base only in style fields — text colour, backing colour, corner radius, position — so it is the same render with different values. What does move a caption charge is which caption surface you render through and, on the subtitle tool, which preset tier you land in. That is a separate decision from the family map above, and it is worth making deliberately rather than by default.