How to add captions to videos (TikTok, Reels, Shorts)

Why every short video needs captions

A lot of people scroll on mute: on the bus, in bed, at work. The same 2019 survey found 25% watch with no sound even in private, and half said captions matter because they often watch without audio (Streaming Media, 2019). Without captions, your hook is a moving mouth.

Captions also help people who do have sound. They follow along faster, catch product names they’d mishear, and stay with you through the parts that would otherwise feel slow. For deaf and hard-of-hearing viewers, they’re the whole video.

Burned-in vs closed captions

For short-form UGC, burned-in wins: it’s part of the look, and it survives reposts, ad accounts and downloads. On YouTube, add an SRT as well.

Built-in auto captions on each app

TikTok

TikTok’s editor has a Captions option before you post. It transcribes your voice and lets you edit the text. Viewers can hide them. The style is fixed and plain, so many creators burn in their own captions and leave TikTok’s off to avoid double text.

Instagram

Instagram launched auto-generated captions in 17 languages in 2022; viewers switch them on or off with a “Show captions” toggle in advanced settings (TechCrunch, 2022). Because viewers control them, don’t rely on them for your hook. Instagram’s Captions sticker in the Reels editor gives you burned-in text with a few styles.

YouTube

YouTube uses speech recognition to create automatic captions on both long-form videos and Shorts (YouTube Help, 2026). YouTube itself warns they may misrepresent speech because of accents, mispronunciations or background noise, and tells creators to always review and edit them in YouTube Studio under Subtitles (YouTube Help, 2026). You can also upload an SRT there.

How to add captions, step by step

1. Get a transcript. Use your editor’s auto-transcribe. Clean audio makes this far more accurate: film close to the mic, away from noise.
2. Fix the words. Read the whole thing. Correct names, numbers and brand terms. Write numbers as digits (“$20”, “3 days”), they read faster.
3. Break into chunks. Two to four words per caption for a punchy look. Break at natural phrase boundaries, never between an adjective and its noun.
4. Time to the voice. Each chunk appears as the first word is spoken and leaves as the next begins. No gaps, no captions lingering over silence.
5. Style and place. Bold sans-serif, white with a dark outline or shadow, big enough to read on a phone. Around 60% down the frame. Check it with the safe zone checker.
6. Emphasize keywords. Color or enlarge one word per line at most: the product, the number, the result.
7. Export and check on a phone, with the sound off.

Caption style that reads in a glance

• Size: bigger than you think. Test by holding the phone at arm’s length.
• Contrast: white text with a black outline or a soft shadow reads on any background. Avoid thin fonts and colored text on busy backgrounds.
• Pace: TikTok recommends 5 to 10 words per second of on-screen text (TikTok Ads Help Center, 2026). Word-by-word captions handle fast talkers naturally.
• Position: one consistent spot. Captions that jump around are tiring. Move them only to stay off the face or out of a demo shot.
• Emoji and callouts: a few, on the words that matter. Not on every line.

SRT files, explained

SRT (SubRip) is the most widely accepted caption file. It’s plain text, so you can open and fix it in any text editor. Each caption is a block: a number, a time range, the text, then a blank line.

1
00:00:00,000 --> 00:00:01,400
If your lunch has ever

2
00:00:01,400 --> 00:00:02,900
leaked all over your bag

3
00:00:02,900 --> 00:00:04,200
this is for you.

• Timecodes are hours:minutes:seconds,milliseconds, with a comma before the milliseconds.
• Blocks are numbered in order, starting at 1.
• Captions shouldn’t overlap: each one ends when the next starts.
• Keep lines short: one or two lines per block.

Need to fix timing or typos in an SRT? Our free SRT editor runs in the browser, and the caption counter checks your post captions against each app’s character limit.

Captions without the manual work

Mimic captions every word automatically, timed to your voice, in the caption style of a reference video you love or one of its built-in styles. Captions stay inside the apps’ safe zones and off your face, and you can change them by typing: “bigger captions”, “make the product name pink”, or fix a word by hand.