Translate video captions on iPhone (19+ languages, on-device)

You made a clip that works. The hook lands, the captions are clean, and the analytics show something you didn’t plan for: a chunk of your viewers are watching from Mexico City, São Paulo, or Berlin, reading captions in a language they only half-follow. Short-form feeds are the most international distribution channel that has ever existed — TikTok and Reels don’t care where your audience lives — but your burned-in English subtitles do. Translating video captions on iPhone is how you stop leaving that audience on the table, and in 2026 it no longer requires a web uploader, a per-minute translation fee, or exporting subtitle files to shuttle between apps.

This guide covers the whole job on-device: auto-caption a video on your iPhone, translate the captions into any of 19+ languages, and export a ready-to-post vertical clip with the translated subtitles burned in — no upload, no account, no subscription.

Why translated captions beat everything else you could do

Before the how, the case for bothering. There are four ways to serve a non-English-reading audience with a short clip, and three of them are bad:

Rely on the platform’s auto-translate. TikTok and Instagram can machine-translate on-screen captions for some viewers, sometimes. Coverage is inconsistent across regions and app versions, the rendering is generic, and the translation vanishes the moment your clip is downloaded and reshared. You’re delegating your international first impression to a feature you can’t see or check.

Post English-only and hope. Muted autoplay is the norm, so captions carry most of the message. A viewer who can’t read them scrolls. You’ll never see this loss in analytics — the clip just quietly underperforms in every geography where it should have worked.

Re-record or dub per language. Actual localization, and actual overkill for a 30-second clip. AI dubbing tools exist on the web, but they’re subscription-priced, upload-based, and a different tier of effort entirely.

Translate the captions, keep the original audio. The pragmatic middle. The viewer hears your real voice — tone, laughter, energy travel fine across languages — and reads subtitles in their own language. It’s the same convention that makes foreign-language film clips work in your own feed, and it costs minutes, not a production cycle.

That fourth option is what this post is about, and the reason it’s newly easy is that Apple ships the translation engine inside the OS.

The on-device way: caption, translate, export in one app

Clipolette is a native iPhone app (iOS 18 or later) that runs the entire pipeline locally: transcription through Apple’s system speech frameworks, translation through Apple’s Translation framework, caption burn-in on your phone’s own GPU. No account, no upload queue, no server that ever sees your footage.

Here’s the full workflow:

Import your video. One file at a time, from Photos or Files — an MP4 or MOV. A clip you shot, a screen recording, a segment your editor sent over AirDrop.
Let it transcribe. Captions generate on your iPhone in the original language. On iOS 26 the SpeechTranscriber pass is guaranteed on-device; either way, nothing uploads.
Fix the transcript first. This step matters more for translation than for ordinary captioning: a misheard word doesn’t just look wrong, it translates wrong. Correct names, brands, and jargon in the editable transcript before translating — Find & Replace with Replace All fixes a repeated mishear across the whole project in one pass.
Translate the captions. Pick a target language from the 19+ that Apple’s Translation framework covers. The first time you use a language, iOS downloads its language pack — a one-time download, and the single network touch in this whole workflow beyond the App Store itself. After that, translation into that language works fully offline.
Style the captions. Word-by-word highlighted animation in five styles — Yellow Pop, Bold White, Cyan Burst, Beast, Minimal — with color and size controls. Translated text can run longer or shorter than the original (German expands, Chinese compresses), so glance over the preview for line fit.
Export. Every clip exports at 1080×1920 — the 9:16 vertical spec TikTok, Reels, and Shorts want — with the translated captions burned into the pixels. The share sheet hands it to Photos or Files, ready to post.

The whole loop for a 60-second clip is a few minutes of attention, and after the one-time language pack download it works on a plane. Install Clipolette from the App Store and run one real clip through it — the translation quality check takes less time than reading this post.

If you’re new to the captioning side of this workflow, the add captions to video on iPhone guide covers transcription accuracy, styling, and platform safe zones in depth; this post assumes that base and adds the language layer.

What “on-device translation” actually means here

The phrase gets used loosely, so here’s precisely what happens. Apple’s Translation framework is the same engine behind the system Translate app. Language models are downloaded and managed by iOS itself — one pack per language, fetched once — and after that, translation runs locally with no per-use network call. Clipolette hands your caption text to that framework and burns the result into the export.

Three consequences worth understanding:

Your content never uploads. Web subtitle translators require your video, or at least your full transcript, on their servers. Here, neither leaves the phone. For client work, prerelease material, or anything sensitive, that’s not a nicety — it’s the difference between “no custody question exists” and “check the vendor’s retention policy.”

There’s no meter. Cloud translation tools price per minute of video or per thousand characters. The Translation framework has no usage fee, and Clipolette doesn’t add one: the app is free with every tool included, and the only purchase that exists is the one-time $24.99 watermark removal.

Quality tracks Apple, not a startup. Machine translation of conversational speech is good in 2026, not perfect. Idioms, slang, and culture-specific jokes are where any engine — cloud or local — stumbles. The mitigation is step 3 above: clean source captions translate dramatically better than sloppy ones, and plain phrasing survives translation better than wordplay.

One clip, several languages

Because output is one exported file per clip, the multilingual workflow is: finish the clip once — trim, reframe, caption, fix the transcript — then translate and export per language. Spanish pass, export, Portuguese pass, export. Each export is a self-contained 9:16 MP4 with that language’s subtitles burned in.

What to do with the set depends on the platform. TikTok’s per-region distribution means many creators simply post the localized version to the same account and let the algorithm route it. Instagram favors the same Reel posted with translated captions burned in over relying on its own caption layer. If you run separate regional accounts, each gets its native-language export. The strategy is yours; the point is that producing the localized files stops being the bottleneck.

Worth stating plainly: this is subtitle translation, not dubbing. The audio stays in your voice and language. Clipolette does no voice synthesis, voice cleanup, or AI dubbing — a deliberate scope choice. For short-form, subtitled-with-original-audio is the established convention anyway; viewers are fluent in reading it.

Translating captions on clips cut from long recordings

Everything above applies to a clip you already have. If your source is a 60-minute podcast, interview, or stream VOD, the translation step slots into the end of a longer pipeline: Clipolette first finds the strongest moments — pick a target length of 15, 30, 60, or 90 seconds and run Find Best Moments, with picks arriving ranked, each with a plain-language reason — then captions each keeper, and then you translate and export each one in the languages you need. On iOS 26 with Apple Intelligence hardware (iPhone 15 Pro or newer) the AI picking runs on-device; on other iPhones you choose scenes manually with keep toggles.

The turn long video into TikTok on iPhone guide covers that extraction pipeline end to end, and the podcast-to-shorts-on-iPhone workflow is the podcast-specific version. A Spanish-subtitled highlight from an English interview is a genuinely differentiated post — almost nobody in your niche is doing it, because until recently it took a web tool and a budget.

Practical quality checklist for translated captions

Fix the source transcript before translating. The single highest-leverage step. Garbage in, mistranslated garbage out.
Keep proper nouns as proper nouns. Check that names and brand terms survived translation untouched; correct any that got “translated” in the editable transcript.
Prefer plain phrasing in the original. If you know a clip is destined for translation, idioms and wordplay in the hook are risk. Plain speech translates cleanly.
Watch line length in expanding languages. German and French routinely run 20–30% longer than English. The word-by-word caption animation handles pacing, but check the preview for crowding and drop the caption size a notch if needed.
Mind the safe zones. Translated or not, captions sitting under TikTok’s UI stack are captions nobody reads. Safe Zones preview in the app shows what TikTok, Reels, and Shorts cover — the vertical cropping guide covers framing in detail.
Spot-check with a native reader when stakes are high. For a client deliverable or a pinned post, two minutes of native-speaker review beats any engine’s confidence.

When this workflow is the right call

You have measurable international viewership — check your analytics for geography — and your clips are caption-dependent, which on muted feeds is all of them.
You post in a niche that travels: cooking, fitness, tech, music, gaming, crafts. Visual-first content plus translated captions is a proven combination.
You handle footage you don’t want uploaded, and every web-based subtitle translator is therefore off the table.
You want the cost to be zero marginal dollars per language, per clip, forever.

When it isn’t

You need dubbed audio in the target language. That’s a different product category — cloud tools do it at subscription prices; Clipolette deliberately doesn’t.
Your target language isn’t in Apple’s Translation framework yet. Coverage is 19+ languages and tracks what Apple ships; check the app’s language list against your needs first.
You need editable subtitle files (SRT/VTT) for a platform’s own caption system. Clipolette burns captions into the pixels — that’s the point for short-form, but it’s the wrong shape for YouTube long-form subtitle uploads.

The bottom line

Your clips already travel internationally; the only question is whether viewers can read them when they arrive. Translating video captions on iPhone used to mean a web uploader, a subtitle-file round-trip, and a fee. Now it’s a one-time language pack download and a few taps in the same app that captions the clip — on-device, offline after setup, free.

Get Clipolette on the App Store, take one clip your analytics say has a foreign audience, and export it with translated captions. It’s free with every tool included — no subscription, and a one-time $24.99 purchase removes the watermark if you want it gone. If the localized version outperforms the original in that geography, you’ve found distribution you were already earning and simply not collecting.

FAQ

Does translating captions on iPhone require an internet connection? Only once per language. Clipolette uses Apple’s Translation framework, which downloads a language pack the first time you translate into a given language. After that download, translation runs on-device and works fully offline — your video never uploads anywhere.

Does Clipolette translate the spoken audio or dub the voice? No. Clipolette translates the caption text and burns the translated subtitles into the exported clip. The original audio stays untouched — there is no AI dubbing or voice synthesis, which is a deliberate scope choice, not a missing feature on a roadmap.

How many languages can I translate video captions into? 19 or more, through Apple’s Translation framework — the same engine behind the system Translate app, covering the major languages you’d expect. Check the language list in the app for current coverage, since it tracks what Apple ships with the OS.