Captions stopped being subtitles

Classic subtitles exist so people can follow along. They sit at the bottom, stay the same size and try not to be noticed. Short-form captions do a different job. Most people scroll with the sound off, so the caption is the hook. It has to read in a glance and give the viewer a reason to stop.

Stacked captions do that better than a single line of text because they add hierarchy. When one word in a block is bigger, brighter or in a different font, the eye goes there first. "This changed EVERYTHING" lands harder than "This changed everything", even though the words are the same.

What makes a stack feel right

1. Break where people breathe

Good blocks end where the speaker pauses or a sentence finishes. Break in the middle of a phrase ("I bought / the house") and the viewer reads two fragments that mean nothing on their own. This matters more than any font choice.

2. One hero word per block, and make it the right one

The big word should carry the meaning: the number, the verb, the surprise. If you make "the" or "and" big because it happened to land on the right line, the whole stack reads as random. One hero word is plenty. Two in a small block cancel each other out.

3. Keep blocks short

Three to six words is the comfortable range for vertical video. Longer blocks turn into paragraphs; shorter ones flicker by too fast to read.

4. Let short words share a line

"So I" or "to be" on lines of their own look like mistakes. Pairing tiny words side by side keeps the stack tight and the shape interesting.

5. A little imperfection helps

Perfectly centred lines can look mechanical. A small sideways offset between lines, alternating or random, makes the stack feel hand-set, like a designer placed it on purpose.

Why it's so slow to do by hand

None of these rules are hard on their own. The problem is volume. A one-minute clip has around 150 words and maybe 40 blocks. For each block you decide the breaks, split the lines, pick the hero word, size it, then nudge everything so it doesn't overlap. Then the client asks for a different font and the line widths change.

That's why many editors give up on the style for anything longer than a 15-second hook, or settle for a template that ignores the content and makes every third word big.

How we approached it in Captyne

We added stacked blocks to Captyne in version 1.5.8, built around the rules above:

Then there's AI Stacking. One click and the AI reads the whole transcript and decides where each block breaks, how its lines split and which word goes big. It follows your block size and your preset's emphasis tags, so the result fits your design rather than a generic template. Don't like a version? Click again for another one. Everything it decides opens in the preview first, so you can change any word before a single layer is built.

After generating, every block stays editable: select the ACP_MASTER null and adjust line padding, jitter and position for the whole video at once.

A few honest tips

Try stacked captions on your next clip

Stacked blocks work on every Captyne plan. AI Stacking is included with Individual Pro, Studio and Enterprise.

See Plans & Pricing →

Keep reading