How to add captions to clips

Why clips need captions

Many people scroll short video with the sound off, in public, in bed, or at work. Without captions, a clip built on what someone says is just a person talking silently.

Captions also help with fast speech, accents, background noise and game audio drowning out a voice. And many clipping campaigns ask for captions in their brief, so check it before you post.

Word-by-word captions

The style most short clips use shows a few words at a time, timed to the speech, often with the current word highlighted. It's easy to follow at a glance and keeps the eye on the screen.

Keep lines short: one to four words on screen at a time is easier to read than a full sentence. Break lines where the speaker naturally pauses, not mid-phrase.

Renafy's AutoClip cuts every clip with word-by-word captions already on. You can restyle them, move them and fix any word in the editor.

Placement: stay clear of the platform's UI

TikTok, Reels and Shorts all put buttons, the username, the post caption and other controls over the video, mostly down the right side and across the bottom. Captions placed there get covered.

Put captions around the middle of the frame, or just below it, with room on both sides. Check each platform's current layout on your phone before settling on a position, because the overlays change from time to time.

On a webcam and gameplay layout, with the facecam on top and the game below, captions often sit well along the line between the two, where they don't cover the face or the key part of the game.

Size, font and contrast

Most people watch on a phone held at arm's length. Captions should be readable without squinting: bold, large and in a simple sans-serif font.

Contrast matters more than colour. White or yellow text with a thick dark outline or shadow stays readable over any background, including bright gameplay and busy scenes. Thin text with no outline disappears the moment the background gets light.

Use one style per clip, and keep it consistent across your account if you can. Viewers come to recognise it.

Keep captions to a size and position that never covers the speaker's face or the action they're reacting to. If the caption hides the moment, move it or shrink it a little.

Fix transcription errors before posting

Automatic transcription is good but not perfect. It stumbles on names, game terms, slang, emotes said out loud, and people talking over each other. A wrong word in a caption looks careless, and a wrong name can change the meaning.

1. Play the clip with the sound on and read along with every caption.
2. Fix names, game terms and slang first: they are the most common errors.
3. Remove filler captions for "um" and "uh" if they add nothing.
4. Check that the caption timing still matches after any trim.
5. Watch it once more with the sound off, as most viewers will.

Timing and line breaks

Captions should appear as the words are spoken, not before or after. Even a small lag feels wrong once you notice it, and it makes punchlines land late.

Let a caption stay on screen long enough to read. When someone talks very fast, show fewer words at a time rather than flashing a long line for a fraction of a second.

Leave a beat of silence uncaptioned. When the streamer pauses before a punchline, an empty moment on screen builds the same tension it does in the audio.

If two people talk, keep their lines in separate captions so viewers can follow who said what. A different colour per speaker can help in longer back-and-forths, but keep it simple.

Emphasis

Highlight the words that carry the moment: the punchline, the shocking word, the number, the name. A colour change or a slightly larger size on one word draws the eye to it.

Use emphasis sparingly. If every line has a highlighted word, none of them stand out.

Don't censor or change what someone said to make it punchier. Captions should match the audio.

Questions