How to Caption YouTube Shorts (Timing That Actually Lands)

Captions for YouTube Shorts have two jobs: stay readable on a phone held at arm's length, and stay out from under YouTube's own interface. Most caption tutorials skip that second part, and it's the one that quietly ruins Shorts, because the app stacks a title, a channel handle, a subscribe button, and a whole rail of action icons on top of your video. This post covers where to put captions on a vertical 1080×1920 frame, how to get word-level timing without guessing, and which caption styles actually hold up at phone size. I shot every screenshot in ClipCraft's Video Captions tool, which runs in the browser and burns the captions into the export.

Why burned-in captions beat YouTube's auto-CC

YouTube already generates captions for Shorts, so it's fair to ask why anyone bothers burning their own in. Three reasons. Auto-CC is something the viewer has to have turned on, so a chunk of your muted viewers never see it. It renders as plain small text pinned near the bottom, in the same neighborhood as your title overlay, and you get no say in that. And it has no rhythm: no word-by-word highlight, no pop, nothing that pulls the eye back to the frame.

Burned-in captions are pixels in the video. They show for every viewer, muted or not, styled and placed exactly where you decided. On a feed where a big share of viewers scroll with the sound off, the caption often does more retention work in the first two seconds than the footage does. That's the case for doing this properly instead of trusting the auto layer.

Where to put captions for YouTube Shorts (the safe zone)

Picture the Shorts player. Along the right edge, roughly the lower half, sits the column of like, dislike, comment, share, and remix buttons. Across the bottom runs your channel handle, the video title, the audio tag, and a progress bar. In practice the bottom 15 to 20 percent of the frame and the right edge of the lower third belong to YouTube, not to you.

That collides with where caption tools naturally put text. ClipCraft's default caption position is 78 percent down the frame, which looks right on a plain preview and then lands under the title overlay once the Short is live. The fix is one slider. Open the Customize panel and drag Position up; the slider runs from 15 to 92 percent of the frame height, and for Shorts I park captions somewhere between 55 and 65 percent. That keeps them below the speaker's face but clear of everything YouTube draws.

Size lives in the same panel and runs from 3 to 11 percent of the frame height. The bold presets default around 6 percent, which works out to roughly 120 px lines on a 1080×1920 export. That sounds huge until you watch it on a phone, where it reads as normal. If you catch yourself shrinking captions so they look tasteful on a desktop monitor, you're designing for the wrong screen.

Timing that actually lands

Timing is what separates captions people read from captions people ignore. You have two options here, one free and one that costs a little.

The free path is tap-to-sync. Type or paste your script (one line per caption, or paste a paragraph and the tool splits it into chunks of 4 words, adjustable from 1 to 10), then hit Tap to sync. The video plays from the start and every tap of Space or Enter stamps the next line's start time while closing the previous one; Esc bails out if you fumble. I ran a five-line script over a 7.9-second vertical test clip and the taps landed at 0.6, 2.2, 3.7, 5.3, and 6.7 seconds, one real-time pass, no typing timestamps by hand. Every row stays editable in 0.1-second steps afterward, so a late tap is a two-second fix.

The paid path is AI auto-captions. The tool listens to your video with Whisper and returns every line with per-word timestamps, so the highlight hits each word at the moment it's spoken. Tap-synced lines time their words by length within the line, which is close; the AI path is exact. It costs 5 tokens per started minute (a free account comes with 100 tokens a month, and token packs start at $0.99 if you run out), the estimate shows on the button before you commit, and a job that fails to start refunds itself.

Styles that survive a phone screen

There are 82 caption presets in the gallery, and being honest, a good third of them are wrong for Shorts. The thin script styles and the small subtitle looks are built for slower, full-screen video. On a phone, over moving footage, they vanish. Here's the same frame twice: Cinema Subtitle, a thin italic serif that sits at 86 percent down the frame, against Beast Mode pulled up to mid-frame.

What the working styles have in common: heavy weight, a dark outline or a solid pill behind the text, and high contrast against anything. For talking-head Shorts, Beast Mode, the Hormozi-style yellow pill, and Podcast Bar (white text on a solid box) are the reliable picks. For gaming clips the gamer pack fits the content, and 8-Bit, Streamer, and Boss Level carry enough outline to survive busy gameplay footage; the pure neon-glow styles look great on dark scenes and smear into invisibility the moment the action turns bright. When in doubt, add outline. The customizer has a slider for it.

If you want the full tour of the gallery, the search box, and the tag chips, the TikTok captions guide walks through all of that; everything in it applies here, minus the placement advice.

Exporting at Shorts size

Shorts want vertical 9:16, and 1080×1920 is the standard. The export keeps your video's resolution up to 1920 on the long side, so a 1080×1920 source comes out exactly 1080×1920, no downscale. The render happens in your browser: the video plays through once in real time while a canvas paints each frame with the captions on it, so a 45-second Short takes about 45 seconds to export. Chrome typically produces a WebM file, and YouTube accepts WebM uploads without complaint.

• Each export costs 2 tokens on the free plan, 1 on Saver, and nothing on Creator or Studio (see pricing).
• Styling, tap-sync, the customizer, and the preview are free with no limits, so you can experiment as long as you like before spending anything.
• The preview draws captions relative to the video frame, not the player box, so what you see on the pillarboxed preview is proportionally what the export renders.

Two related workflows, briefly. If the clip needs cutting before it needs captions, do that first; our guide to editing AI-generated video covers the trim-and-composite pass in VideoCraft. And if your Short is a song rather than talking, the tool's Create-from-audio mode builds a 1080×1920 video from an audio file plus a background, captions included; the Suno music video post shows that whole pipeline.

My honest recommendation: pick one bold style and stop auditioning, put it at 60 percent, and spend your saved time on timing, because viewers forgive a plain caption long before they forgive one that's covered by a subscribe button. A free account covers a full test run, AI timing included.