How to add captions to YouTube Shorts
Many people watch Shorts with the sound low or off, and a caption is often what stops them swiping. There are three ways to add one, and they behave very differently. This guide explains each, when to use it, and how to make captions people can actually read on a phone.
Published
See a supplied-caption render
The following example uses an original synthetic source and supplied caption timings. It shows an actual rendered file, not an automatic-transcription benchmark.
1. YouTube's automatic captions
YouTube creates automatic captions for many languages after you upload. They cost nothing and need no editing software, but viewers only see them if they turn captions on, and they all look the same: small white text in YouTube's own style.
Check them before you rely on them. Automatic captions often get names, brand words and mixed-language speech wrong. You can correct them in YouTube Studio under Subtitles.
2. Captions burned into the video
Burned-in captions are part of the picture, so everyone sees them, whether or not captions are switched on. This is what most popular Shorts use: big, bold words in the middle of the screen.
The cost is time. Typing and timing captions by hand in an editor takes far longer than the clip itself, and once the video is exported you can't fix a typo without exporting again.
3. AI word-by-word captions
AI caption tools transcribe the speech and burn in captions for you, timed to each word. The popular style highlights each word as it is said, which keeps the eye moving with the voice.
Pick a tool that lets you correct the transcript before the video is rendered, and one that understands the language you actually speak, including a mix like Hindi and English.
Make captions readable on a phone
- Short lines: two to four words at a time. Long sentences across the screen are read slower than they are spoken.
- Large and bold, with an outline or a box, so they stay readable over any background.
- In the middle of the frame, a little above or below centre. The bottom and the right-hand edge are covered by the Shorts player's title, buttons and channel name.
- Never over a face or over text that is already in the video.
- Spelled right. A misspelt name is the first thing commenters notice.
Which one should you use?
Turn YouTube's automatic captions on for every upload: they help viewers who rely on them and cost nothing. For the captions most viewers will see, burn them in, either by hand for a single clip or with an AI tool when you post regularly.
How TheMomentra handles this
TheMomentra transcribes your video in English, Hindi, Marathi or a mix, flags the words it wasn't sure about and lets you correct them before anything is rendered. Captions are burned in word by word in one of 12 styles, placed in the upper, centre or lower safe area. The Free plan includes 30 minutes of video a month.