Long sentences
One word at a time is a pace, and a pace has a ceiling. Dense explanation delivered fast leaves the viewer behind; a karaoke line they can read ahead of handles that better.
Caption style
Every guide to this look tells you to pick a heavy sans and highlight the key word. That is the shape of it, not the thing itself. Here is the actual preset, down to the hex, and a generator that applies it to your own footage.
Hormozi — Acid-green caps; the key word punches out white under a heavy shadow.
Closest alternatives — point at one to play it
Hormozi Pop
Karaoke
Beast Mode
Two presets carry this look. Both are free to preview on your own clip before you export.
The effect people are copying is not the yellow, and it is not the font. It is that the caption never shows you a finished sentence. Words arrive one at a time, on the beat of the speech, so there is nothing to read ahead to — your eye has no choice but to stay with the audio. That is what makes it feel fast on a silent autoplay.
The colour choice is doing a second job. A saturated resting colour with a white active word inverts the usual relationship: normally the highlight is the bright thing. Here the whole line is bright and the spoken word is the calm one, which is why it stays legible over footage that is already loud.
The 3px stroke is the part most recreations get wrong. Without it the type dies against any light background, and at 4px or more the counters in Montserrat 900 start to fill in and the caps turn into blocks. Three is the number that survives a white wall and still reads as type.
One word at a time is a pace, and a pace has a ceiling. Dense explanation delivered fast leaves the viewer behind; a karaoke line they can read ahead of handles that better.
Acid green on near-black is a strong opinion. If your video sits inside a brand system, change the resting colour in the editor — the motion is the part worth keeping.
The look was built for high-energy business content and it carries that register with it. It will make a calm piece feel like an advert.
Word-level motion puts every single word on screen alone, which means every transcription error is isolated and obvious. Names, figures and product terms need a read-through before export.
Bring in the edited short, or the raw take if you plan to cut after.
It sits in the Creators family. The preview plays your own footage, not a sample.
Fix names and numbers, then export a captioned video or a subtitle file.
The look is built on a heavy geometric sans in all caps. This preset uses Montserrat at weight 900, which is the closest widely available match and the face most commonly used to recreate it. The weight matters more than the specific family: below about 800 the style stops reading as this look at all.
Both are in circulation, which is why recreations disagree. This preset rests the line in acid green and punches the spoken word white. The other preset on this page, Hormozi Pop, does the more familiar thing: white caps with the spoken word turning gold. Preview both and pick by your footage — the green holds up better over bright frames.
They are synced to the speech. Transcription returns word-level timestamps and the renderer uses those spans directly, so a word held for three quarters of a second stays lit for three quarters of a second. Where a clip comes in without usable word timings, the words are divided evenly across the cue instead, which keeps the motion but loses the exact sync.
Yes. The preset is a starting point — resting colour, highlight, outline weight, size and vertical position are all editable, and the change shows on your own frame immediately rather than on a sample clip.
No. You can upload a clip, generate captions and preview any style without paying. Exporting a finished video is what consumes minutes.
The read-ahead alternative when the script is dense rather than punchy.
Word-level motion hides bad line breaks until the moment it does not.
Why 80% of frame height, and what covers the caption on each platform.
The same workflow, framed for a vertical TikTok edit.
Upload a clip, apply the preset, read the transcript back, and export.