There are three ways to put words on a Reel, and the internet keeps writing tutorials for the first one as though the other two don’t exist. They do different jobs, they fail differently, and picking wrong is why people end up hand-timing text boxes at midnight. Here’s all three, honestly, including the parts Instagram’s own tool can’t do.
Route 1 — Instagram’s captions sticker
The built-in one. It takes about half a minute:
- Record or upload your clip and tap Next into the Reel editor.
- Tap the sticker icon in the top bar.
- Choose Captions. Instagram transcribes the audio and drops in timed text.
- Tap the text to fix mis-transcribed words, switch between the handful of built-in styles, and change the highlight colour.
- Drag the block where you want it, then share.
For a single Instagram-only clip this is genuinely the right answer, and you should use it. Its limits are worth naming precisely, because they’re structural rather than cosmetic:
- It doesn’t exist in your file. The sticker is rendered by Instagram, on Instagram. Download that Reel to post on TikTok and the captions are gone — you do the whole job again in a different app with a different look.
- A fixed set of styles. You get the ones Instagram ships. No custom typeface, no per-word colour, no placing one word behind your head.
- Sentence-level timing. Corrections are per-word but timing generally isn’t, so a clip with music, an accent or overlapping speech tends to drift — and the drift compounds down the clip.
- No safe-zone awareness. Nothing stops you parking the text under the buttons or in the band your profile grid crops off.
Route 2 — Manual text boxes
The text tool plus the timeline slider: add a text box, set the in and out points, repeat. Total control, and the cost is exactly what it sounds like — a 45-second clip is 25 to 40 lines, each individually timed. Fine for a five-line skit, unsustainable as a habit.
The one case where it genuinely wins: when the text isn’t the transcript. A punchline landing three seconds after a beat, a label pointing at something in frame, a running joke in the corner — that’s hand-placed text and it should be.
Route 3 — Burned-in captions from an editor
Burned-in (also called open) captions are baked into the pixels before you upload. Every serious short-form account works this way, for four reasons that stack:
- One file, every platform. The same export goes to Reels, TikTok and Shorts and looks identical on all three.
- Word-level timing. Each word carries its own timestamp, so the active word can light up in sync with your voice — the single biggest retention upgrade available, and the thing every style in our field guide is built on.
- Real typography. Your typeface, your colour, your emphasis rules, consistent across every post — which is how a feed starts looking like a brand rather than a person posting.
- Effects that the sticker can’t do at all. Text passing behind the speaker, per-word scaling, highlight boxes that follow the beat.
Rules that apply whichever route you pick
- Three or four words a line. Past that it stops being a glance and becomes reading, and reading competes with listening instead of reinforcing it.
- Stay out of the interface. The top band holds the account row, the bottom holds the caption tray and buttons. Keep text in the middle. The same band survives the profile-grid crop — see the safe-zone guide.
- Contrast is not optional. White on a bright kitchen wall is invisible. An outline, a shadow or a filled box behind the words costs nothing and fixes it permanently.
- Emphasise one word per phrase, not all of them. Emphasis works by contrast. If everything is big and coloured, nothing is.
- Say your keywords out loud and show them. Platforms transcribe your audio to work out what the clip is about. When the spoken line and the on-screen line match, both signals point the same way.
Three problems, and what actually causes them
| Symptom | Real cause | Fix |
|---|---|---|
| Captions drift later and later | Sentence-level timing, or a trim applied after generation | Word-level timestamps; re-time after trimming, not before |
| Words are wrong on names and jargon | The transcriber has no vocabulary for your niche | Fix them in the transcript once — every caption updates |
| Text vanishes on some footage | No outline or backing; white text on a bright frame | Outline, shadow, or a box; or pick a higher-luminance colour |
Doing it in Moonshot
Moonshot is route three with the manual work removed. Upload a clip, and it transcribes every word with its own timestamp, then renders burned-in captions in whichever of the template styles you pick — you can flip between them live without re-editing anything.
The part that saves the most time is that you edit words, not a timeline. Fix a mis-transcribed name in the transcript and every caption clip re-flows around it. Delete a sentence and the video cuts with it. There’s no per-line in and out point to drag, because the words already know when they happen.
Quick answers
How do I add automatic captions to an Instagram Reel?
In the Reel editor, tap the sticker icon and choose the Captions sticker. Instagram transcribes your audio and drops in timed text you can restyle, reposition and correct word by word. It's free, it's built in, and it takes about thirty seconds — it just can't leave Instagram with you.
What's the difference between the captions sticker and burned-in captions?
The sticker is a layer Instagram renders on its own surface: it lives inside the app, offers a fixed set of styles, and doesn't exist in the file you'd upload elsewhere. Burned-in captions are baked into the video pixels before upload, so they look identical on Reels, TikTok and Shorts, and you control the typeface, colour, timing and placement completely.
Why are my Reel captions out of sync?
Almost always because the transcript is timed at the sentence level rather than the word level, so one bad boundary shifts everything after it. Trimming the clip after captions were generated does the same thing. Word-level timing survives edits because each word carries its own timestamp.
Do captions actually help reach?
They help in three separate ways: most feed viewing starts muted so captions are the only thing communicating, they hold the eye on a moving target which raises watch time, and the platform reads your on-screen text and audio to work out what the clip is about. That last one is why saying and showing the same keywords matters.
Should I put the caption in the middle of the frame?
Centre horizontally, and vertically somewhere between the middle and the lower third — but above the bottom band where the username, caption tray and buttons sit. Text in the top or bottom sixth of a 9:16 frame competes with the interface and gets cropped out of your profile grid entirely.
Caption a clip in about a minute
Upload once — Moonshot transcribes every word, times it, and burns in captions you can restyle instantly. The same file works on Reels, TikTok and Shorts.
Open Moonshot free