Add captions to video on iPhone (AI, on-device, 2026)

Roughly 85% of social video plays muted. That single statistic is why captions stopped being optional. A clip without on-screen text gets thumbed past before the audio ever loads — the viewer never learns whether your 30 seconds was worth their attention. On TikTok, Reels, and Shorts, captions aren’t an accessibility nicety. They’re the difference between a watch and a scroll.

The problem is that adding captions to a video on iPhone is still annoyingly fiddly for most people. You either retype the transcript by hand in a video editor, fight with a clunky web uploader, or settle for a tool that mangles every proper noun and forces a subscription that scales with how much you post. This guide walks through how to add accurate, burned-in captions to a video directly on your iPhone — on-device, offline, and without uploading your footage anywhere.

Why “add captions” is harder than it sounds

On the surface it’s a solved problem: transcribe the audio, put the words on screen, done. In practice there are four jobs hiding inside it, and most tools only do one or two well:

Transcription accuracy. Auto-captions that mishear every name, brand, and technical term create more cleanup work than typing from scratch. The first 90% is easy; the last 10% — proper nouns, fast speech, overlapping talkers — is where tools fall down.
Timing. Captions have to track speech tightly. Words that lag or lead by even a quarter-second feel broken and pull attention away from the content.
Legibility. Text has to read on a 6-inch screen in bright sunlight, over busy footage, without covering the speaker’s face. That’s a styling problem, not just a transcription one.
Burning in. Platforms like TikTok and Reels show their own auto-captions inconsistently, and they vanish when someone downloads or reshares your clip. Captions baked into the pixels travel everywhere the video goes.

Doing all four by hand, per clip, is the reason most creators either skip captions or post inconsistent ones. The fix is a tool that handles all four on the device you already shot the video on.

The fast way: caption on-device with Clipolette

Clipolette is a native iPhone app (iPhone 15 Pro or later) that runs the entire transcription-and-caption pipeline on the device’s Neural Engine. There’s no upload, no queue, and no internet requirement — the speech model ships inside the app. You hand it a video, it transcribes the audio, burns in captions, and exports a finished file straight to your Camera Roll.

The full workflow takes about a minute of your attention:

Open Clipolette and tap Import. Pick a video from Photos, the Files app, or iCloud Drive. Nothing uploads — the file is read locally.
Choose your format. 9:16 for TikTok / Reels / Shorts, 1:1 for an Instagram feed post, or 16:9 to keep it landscape.
Let it transcribe. The Neural Engine indicator lights up. A 60-second clip captions in a few seconds; a multi-minute clip takes proportionally longer but still runs entirely on-device.
Review and fix inline. The transcript appears synced to the video. Tap any word to correct it — this is where you fix the names, brands, and jargon that every auto-caption tool guesses wrong.
Pick a caption style. Clipolette’s default is tuned for muted-feed legibility: high-contrast text, sensible line breaks, positioned to clear the speaker’s face.
Export. The captioned video lands in your Camera Roll, ready to post from the TikTok, Reels, or Shorts app.

Because the audio never leaves your iPhone, this works on a plane, on the subway, or anywhere with no signal — and it works for footage you’d never want on a stranger’s server: client work, NDA’d interviews, anything unreleased.

Install Clipolette from the App Store and caption one clip — you’ll know inside two minutes whether the accuracy clears your bar.

Captions on a clip vs. captions on a long video

There are two distinct jobs people mean by “add captions on iPhone,” and the right approach differs:

You already have a short clip. Maybe you trimmed a 40-second moment in the Photos app or shot it directly. You just want clean burned-in text. Import it, transcribe, fix names, export. That’s the flow above, start to finish in under two minutes.

You have a long recording and want captioned shorts out of it. This is the podcast / interview / stream case. Here, captioning is one step inside a larger loop: the app first picks the strongest moments from the long file, cuts them into individual clips, then captions each one. Clipolette does both — point it at a 60-minute file and it returns a folder of captioned, ready-to-post shorts rather than one captioned monolith. If that’s your situation, the long-form to short-form workflow and the turn long video into TikTok on iPhone guides go deeper.

Knowing which job you’re doing saves time. Don’t run a 90-minute file through a “caption this clip” tool, and don’t hand-trim thirty clips before captioning when the app can find them for you.

The manual alternatives, and where they cost you

For comparison, here’s how people add captions on iPhone without a dedicated AI app, and the tradeoff each carries:

iOS Live Captions / built-in dictation. These are real-time accessibility features, not a way to produce burned-in captions on an exported video. They don’t write text into the pixels, so the captions don’t travel with the file. Useful for watching, not for publishing.

CapCut and similar editors. CapCut has auto-captions and they’re decent. The catch is that the free tier nudges you toward an account and cloud features, the styling can feel template-y, and correcting misheard proper nouns means hunting through a timeline. It’s a capable editor doing captioning as a side feature.

Web uploaders. Plenty of browser tools will caption a video. On iPhone that means uploading your file over cellular or Wi-Fi, waiting in a queue, then downloading the result back into Photos. Slow, dependent on signal, and your footage sits on someone else’s server.

Manual subtitle entry. Typing captions by hand in any editor gives you total control and total tedium. Nobody sustains it past a few clips.

The on-device approach wins specifically when you post often, care about accuracy on names and jargon, or handle footage you don’t want uploaded. If you caption one video a year, the built-in tools are fine.

Getting accurate captions, not just fast ones

Speed is easy; accuracy is the part that actually saves you time. A few habits that help regardless of tool:

Fix proper nouns once, immediately. Names, brands, product names, and acronyms are where every speech model guesses. Correct them in the review pass before export — re-editing a baked-in video later means re-exporting the whole thing.
Record clean audio. The single biggest lever on caption accuracy is the input. A lapel mic or even AirPods beat the iPhone’s built-in mic in a noisy room, and cleaner audio means fewer corrections downstream.
Keep lines short. Two or three words per line reads far better on a phone than a full sentence crammed across the screen. A good caption tool breaks lines for you; if yours doesn’t, break them manually.
Position above the lower third. TikTok and Reels overlay their own UI — the username, caption text, and buttons — across the bottom of the frame. Captions sitting too low get covered. Clipolette’s default placement already accounts for this.

Captioning for each platform

The same clip often goes to more than one place, and each surface treats captions a little differently. Knowing the quirks up front saves a re-export:

TikTok. The right-side button stack and the bottom caption block eat a surprising amount of the frame. Keep burned-in text in the upper-middle third, and don’t rely on TikTok’s own auto-captions — they’re inconsistent and disappear the moment someone downloads or stitches your video. Baked-in text is the only kind that survives a reshare.

Instagram Reels. Similar safe-zone math to TikTok, but Reels crops more aggressively at the top for some feed placements. Center your captions vertically rather than hugging an edge, and they’ll survive every placement Instagram throws the clip into.

YouTube Shorts. Shorts shows a title and channel handle along the bottom. Burned-in captions positioned above center read cleanly, and because Shorts viewers often have sound on, captions here are reinforcement rather than the whole message — but they still lift retention.

LinkedIn. If you’re repurposing a clip square (1:1) for a LinkedIn feed post, captions matter even more: LinkedIn autoplays muted with no built-in caption layer at all. Burned-in text is the entire reason the clip works in that feed.

Because Clipolette exports to your Camera Roll in whichever ratio you pick, you can run the same source twice — once 9:16, once 1:1 — and get correctly-captioned versions for each destination without re-transcribing. The transcript is reused; only the framing and layout change.

Editing captions without re-doing everything

The most underrated part of a good captioning flow is how cheap it is to fix a mistake. With a hand-typed or web-uploaded workflow, catching a wrong name after export means going back into an editor, finding the frame, retyping, and re-rendering the whole video. That friction is why so many published clips ship with an obvious mishear in them — the creator saw it and decided it wasn’t worth the redo.

On-device editing changes the math. In Clipolette the transcript stays editable right up to export, synced word-for-word with the video, so correcting “Sonia” that came through as “Sonya” is a single tap and the caption updates everywhere it appears. Fix it before you export and the finished file is clean the first time. That’s the practical reason on-device captioning compounds: the cost of getting it right drops low enough that you actually do it on every clip instead of just the important ones.

When on-device captioning is the right call

You’re on iPhone 15 Pro or later — the on-device speech and caption models need the Neural Engine in those chips.
You post short-form regularly and want captioning to take seconds, not a manual editing session.
You handle sensitive footage — client work, interviews under NDA, unreleased material — that you’d rather not upload to a web tool.
You shoot or edit on the go and want the whole loop to work without a connection.

When it isn’t

You’re on an older iPhone without the required silicon — the on-device pipeline won’t run.
You need captions in a language the on-device model doesn’t cover yet — check the App Store listing for the current language list.
You want heavily animated, word-by-word karaoke styling as a brand signature — Clipolette’s default is clean and legible rather than maximally flashy, by design.

The bottom line

Adding captions to a video on iPhone should take seconds, produce text accurate enough that you fix two words instead of twenty, and never require uploading your footage to anyone. On Apple Silicon iPhones that’s now possible entirely on-device. The win isn’t just speed — it’s that captioning stops being the chore that makes you skip it, which means more of your clips actually get the on-screen text that earns the watch.

If you want to caption a clip or pull captioned shorts out of a long recording without a queue or an upload, get Clipolette on the App Store — it’s free to use, with every editing tool included and no subscription; a one-time $24.99 purchase removes the export watermark if you want it gone. Caption one real clip, check the accuracy on the names that matter to you, and you’ll know whether it’s earned a place in your posting routine.