Classic captions are speech with a stroke, not a hook line

The Classic group is kind=subtitle: TikTok Sans, white fill, black outline, bottom. It burns what was said. Headline is a different group.

Classic is the everyday social caption. It is not a title card. It is not a karaoke pop. It is not a line you authored because the take had no speech. Pick it when someone already talked and you need that talk readable on mute. Pick something else when the first frame has to be the hook.

kind=subtitle, not a hook

Classic's kind is subtitle. The whole line appears together, timed to the voice. It updates as a block. It does not scale one token. It does not typewriter the sentence. Word-timed families — Snap and Reveal — are kind=timed and they own a different clock. Classic does not have an effect field. There is nothing to pop.

That is the job boundary. Add captions to a video transcribes audible speech and burns styled captions. Speech in; type out. Classic is the look you reach for when that transcript should read as "this is the show" rather than as a poster. Give the captioner a silent b-roll plate and you have spent a pass on a file that had nothing to transcribe. Overlay the line you wanted, or record speech, then caption.

A hook you typed — "WAIT FOR IT," "3 reasons," a price — is not speech. Add a text overlay to a video burns caption_text you supply, whole clip, top / center / bottom. Running Classic on a music-only file hoping the CTA appears will not typeset your copy. Classic only has what was said.

Burn-in is not a generate. Do not prompt a video model to "add white subtitles with a black outline." Scene models paint glyphs. They do not time a transcript. Generate the picture. Caption the locked file.

TikTok Sans, white fill, black outline, bottom

The base preset is Classic (tiktok_classic). Face is TikTok Sans (fontId: tiktoksans). Fill is white. Outline is black, width 1. Position is bottom. That is the everyday social caption: a thin stroke that holds on ordinary talking-head footage without a card.

TikTok Sans here is a caption/overlay face. It is not an image model. It is not Ideogram. You are not generating lettering into the plate. You are burning a transcript in a named font.

Bottom is the default because speech captions live in the lower third. If the subject already occupies that band — a product in hand, a lower-third name, a burn you already printed — Classic will collide. That is a layout problem, not a reason to switch families. Use the top variant, or move the overlay, or recrop. Do not jump to Headline because the lower third is busy. Headline is a different group with a different job.

Classic is the family for footage that is busy but not hostile: a real room, a real shirt, a real wall. A thin black ring is enough. Gym floors, weather, and hard cuts are Halo or Hype. Cards are Paper or Ink. Classic is the stroke.

The five variants stay inside Classic

The group is five looks. They change colour and position. They do not change kind, face, or the fact that this is speech.

• Classic (tiktok_classic) — white fill, black outline, bottom. The one to judge the family by.
• Classic Yellow (classic_yellow) — fill #FFE94A, black outline, bottom.
• Classic Mint (classic_mint) — fill #7CF5C4, black outline, bottom.
• Classic Navy (classic_navy) — white fill, outline #0B1F44, bottom.
• Classic Top (classic_top) — white fill, black outline, top.

Yellow and mint are still subtitles. They are not a hook family. Navy is still a stroke, not a card. Top is still Classic: same face, same kind, different band. Pick one variant and hold it. Switching Classic Yellow on Monday and Classic Navy on Tuesday throws away the recognition the stroke is for.

A house style is this family plus one variant, across the talking clips. This page is only Classic.

Headline is a different group

Headline is tall condensed Bebas, title-card energy, default top. It is still kind=subtitle in the catalog — the line arrives as a block — but the job is the first frame, not the transcript of the take. Recaps, numbers, news-style opens. If the type is the open, you are in Headline. If the type is what the speaker just said, you are in Classic.

Do not use Classic as a fake Headline by typing a hook into a captioner. The captioner will transcribe the take, not your hook. Do not use Headline as fake Classic by setting Bebas on a full transcript. The face will shout every word.

Snap pops each spoken word. Reveal builds the line live. Those are timed groups. Classic will not become them if you "make it punchier." Punch is a different kind.

The test: is there audible speech you want burned on, in a thin stroke, at the bottom (or top) of a talking plate?

Yes: Classic. Name the variant. Caption last.

If the words were never spoken, overlay. If the words are the poster, Headline. If there is no picture yet, generate first.

FAQ

Is Classic a word-timed karaoke look?

No. kind=subtitle. The line is a block. Snap and Reveal are the timed groups. Classic has no effect to pop.

Can I use Classic on a silent product plate to print the pack name?

No. Classic burns what was said. A silent plate has nothing to transcribe. Overlay the name, or record speech and then caption.

What is the default face and position?

TikTok Sans, white fill, black outline, bottom, on tiktok_classic. Classic Top is the same face at the top of frame. The font page is TikTok Sans.

Why not Headline if I want the captions to feel bigger?

Headline is a different group: Bebas, title-card energy, default top. Bigger speech is still speech. Stay in Classic, or change the crop. Switching families because you wanted a heavier stroke is how a transcript starts looking like a poster.