How do I add word by word captions to a video?

Use a tool that has word-level timing from the transcript and a style that shows or highlights one word at a time. In KeyMo, import the video, let it make the clips, then choose a style such as Pulse, Spotlight or Reader.

Are word by word captions better than full-sentence captions?

They tend to suit short, punchy clips. For dense explanations, full lines are easier to read and re-read. Pick per clip, not per show, and keep to one or two short lines either way.

Can YouTube show captions word by word?

YouTube's help page says uploaded WebVTT files support positioning but only bold, italic and underline styling, and it does not describe a per-word animated look. To get that look, burn the captions into the video before upload.

Can Instagram do word by word captions?

The Meta pages we read cover Facebook's Reels composer, which can transcribe audio into an editable caption, and the Instagram Reels ad specs. Neither describes per-word animation. For that effect, add burned-in captions before you upload the clip.

Is there a generator for word by word captions?

Yes. KeyMo is one: it makes the clips and captions each in a word-by-word style, in 12 looks. It has no free plan, so a free generator would be a different tool, and you should read its limits.

How do I keep one-word captions readable?

Use a heavy, plain typeface with an outline or box, keep text away from the app's buttons, and mind contrast: WCAG's general minimum for text is 4.5:1. Then watch the sample on a phone before you export all clips.

Guide

Word by word captions: what they are and how to add them

By Rakshit · Updated October 11, 2026

Word by word captions show or highlight speech one word at a time, in step with the voice, instead of putting a whole sentence on screen. They tend to suit short, punchy clips, while dense explanations read better as full lines. KeyMo times every caption word by word and offers the look in 12 styles, including Pulse, Spotlight and Reader.

Two ways to do it

People use the phrase for two different looks, and the choice changes how the clip feels. In the first, a single word or a very short group appears, replaces itself, and the screen is never crowded. In the second, a full line sits on screen and a highlight moves along it. Someone asking for a TikTok caption that shows one word at a time is probably after the first look.

Three KeyMo styles that run word by word
Look	What the viewer sees	KeyMo style
Words appear one by one	Each word pops in as it is said, in a small group that is replaced by the next	Pulse
Full line with a travelling highlight	The whole line is visible and the spoken word is picked out in a colour	Spotlight
Full sentence that fills in	A sentence on a light box where words darken as they are spoken	Reader

For the second look in its most dramatic form, see karaoke style captions.

When word by word beats full sentences

Short, strong lines. A host delivering a clean hook in a few words keeps pace with the type.
Fast talkers. Small groups of words follow the rhythm, where a full sentence would arrive before the voice does.
Muted viewing. The captions are the only thing carrying the clip, so keeping time with the voice helps.
But not for dense or technical speech. A viewer who wants to read ahead, or re-read, gets less from one word at a time. A full line is easier to scan, and W3C's guidance leans to one or two lines, generally under 32 characters per line.

A simple rule: if you can say the clip's point in one breath, go word by word. If the point needs a clause and a caveat, show the whole line.

What YouTube, Instagram and TikTok offer

Platform captions and per-word looks
Platform	What the vendor's own pages say	What that means
YouTube	Automatic captions exist for long-form videos and Shorts and should be reviewed. Uploaded WebVTT files support positioning, but styling is limited to bold, italic and underline	Nothing on those pages describes a per-word animated look. For that, burn the captions into the video
Facebook Reels and Instagram Reels ads	Facebook's Reels composer can transcribe a reel's audio into a caption with editable font and colour. Meta's ad guide calls on-video captions optional but recommended for Instagram Reels ads	The pages we read do not describe per-word animation, so use burned-in captions for that look
TikTok	Creators choose auto captions on the editing page, transcribed text is editable, and viewers can hide it	The pages we read do not describe a word-by-word look either

These are what the pages say, not a test of every app version. Features change, so check the app before you rely on a setting. The captions on TikTok guide covers the TikTok options in more detail.

How to add them yourself, and the slow part

By hand, every spoken word needs its own start and end time, and each one needs a text layer or a keyframe in your editor. That is why most people use a tool that already has word timings from the transcript. The timing is the hard bit; the look is a few settings.

A caption file can hold some of this: the WebVTT specification lets a cue carry timestamps inside its text that mark when part of the line becomes active. But a player has to render that, and YouTube's help page does not list it as supported, so the dependable route for a social clip is to draw the words into the picture.

How to get them automatically on every clip

Bring in the long video. Paste a YouTube link on a paid plan, or upload an MP4 or MOV between 5 minutes and 3 hours.
Choose the length range: Auto 15 to 90 seconds, or Short, Medium or Long.
KeyMo finds up to 15 clips, scores each out of 10 and keeps those at 7 or higher.
Pick Pulse, Spotlight or Reader, or any of the other styles, and export in 9:16 at 1080x1920 with no watermark.

Captions in KeyMo are word-by-word whichever style you choose. The style decides how each word looks, not whether the timing is per word. If one word is wrong, open the clip in the chat editor and tell KeyMo the fix. Starter is $12 a month for 300 source minutes; there is no free plan. The podcast to Shorts page shows the whole flow.