Translate video captions on Mac (19+ languages, offline)

You cut clips on a Mac because that’s where the source files live — the podcast export from Logic, the Riverside download, the two-hour interview sitting in a project folder. The clips work, and then the analytics hand you a surprise: a real slice of your viewers is in São Paulo, Mexico City, or Paris, squinting at English captions on a muted feed. Translating video captions on Mac is how you collect that audience instead of leaking it, and the usual objections — a web uploader, a per-minute translation fee, an SRT round-trip between three tools — no longer apply. The translation engine now ships inside macOS, and a native app can run the whole job locally.

This guide covers the full pipeline on the Mac: auto-caption a video, fix the transcript, translate the captions into any of 19+ languages, and export a ready-to-post vertical clip with the translated subtitles burned in. No account, no upload, no subscription, and after a one-time language pack download, no internet.

Why the Mac is the right place to do this

You can translate captions on an iPhone — the iPhone version of this workflow covers that — but if your clips start life as long recordings on a Mac, doing the translation there too has concrete advantages:

The source is already local. Web subtitle translators want your video uploaded. A 90-minute 1080p interview is a couple of gigabytes; on the Mac the file never moves, so the upload tax is zero regardless of file size.

Transcript cleanup is faster with a keyboard. Translation quality is downstream of transcript quality, and fixing a transcript is real text editing — exactly the work a Mac keyboard and Find & Replace (⌘F) are built for. Correcting a guest’s name across forty captions is one Replace All, with Undo behind it if you overreach.

Exports land in your folder tree. Each translated clip saves through the standard macOS save panel into whatever folder structure your post-production uses — per-language folders, per-episode folders, however you organize. No share-sheet hop, no downloads folder archaeology.

It’s the same machine that cut the clip. If you’re already extracting highlights from a podcast or webinar on the Mac, translation is one more pass in the same app, not a handoff to another tool.

The on-device workflow, step by step

Clipolette is a native Mac app (macOS 15.4 or later, Intel Macs included) that runs the entire pipeline locally: transcription through Apple’s system speech frameworks, translation through Apple’s Translation framework, caption burn-in on your own hardware. No login, no backend, no server that ever holds your footage.

Import your video. One file at a time, MP4 or MOV, from Files or Photos. A finished clip, a screen recording, a segment exported from Final Cut or Premiere.
Let it transcribe. Captions generate on the Mac in the original language. Nothing uploads at any point.
Fix the transcript before translating. This is the highest-leverage step in the whole workflow. A misheard word in ordinary captioning looks wrong; in translation it translates wrong, and the error compounds. Correct names, brand terms, and jargon in the editable transcript — Find & Replace (⌘F) with Replace All fixes a repeated mishear across the entire 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, macOS downloads its language pack — a one-time download, and the single network touch in this 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 runs longer or shorter than the original (German expands, Chinese compresses), so scrub the preview for line fit before export.
Export (⌘E). Every clip exports at 1080×1920 — the 9:16 vertical spec TikTok, Reels, and Shorts share — with the translated captions burned into the pixels, saved through the save panel into the folder you choose.

The loop for a 60-second clip is a few minutes of attention. Install Clipolette from the App Store and run one real clip through it — checking the translation quality on your own footage takes less time than reading this post.

If the captioning half of this is new to you, the add captions to video on Mac guide covers transcription accuracy, styling, and safe zones in depth; this post assumes that base and adds the language layer.

What “on-device translation” means on macOS

Apple’s Translation framework is the same engine behind the system Translate app. macOS downloads and manages one model 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.

The consequences are worth spelling out:

Your footage never uploads. Web subtitle translators need your video, or at minimum your full transcript, on their servers. Here neither leaves the Mac. For client work, NDA interviews, or prerelease material, that’s not a nicety — it removes the custody question entirely instead of answering it.

There’s no meter. Cloud translation is priced 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.

It works on Intel Macs. Because translation is an OS framework available from macOS 15.4, the caption-translate-export pipeline runs on Intel machines too. The one thing an Intel Mac can’t do is the AI clip picking — more on that below.

Quality tracks Apple, not a startup. Machine translation of conversational speech in 2026 is good, not flawless. Idioms, slang, and culture-specific jokes are where every engine stumbles, cloud or local. 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, one folder each

Output is one exported file per clip, so the multilingual loop is: finish the clip once — trim, reframe, caption, fix the transcript — then translate and export per language. Spanish pass, ⌘E, save into the Spanish folder. Portuguese pass, ⌘E, Portuguese folder. Each export is a self-contained 9:16 MP4 with that language’s subtitles burned in, and the save panel means the sorting happens at export time instead of in a cleanup session later.

What you do with the set is a distribution decision. TikTok’s per-region routing means many creators post localized versions to the same account and let the algorithm sort it out. Instagram rewards a Reel with translated captions burned in over one that leans on its own caption layer. Separate regional accounts each get their native-language export, AirDropped to the phone or uploaded from the Mac through the platforms’ web upload pages.

To be plain about scope: this is subtitle translation, not dubbing. The audio stays in your voice and language — Clipolette does no voice synthesis or AI dubbing, deliberately. For short-form that’s the established convention anyway; subtitled-with-original-audio is how foreign-language clips already work in your own feed.

Translating clips cut from long recordings

Everything above applies to a clip you already have. If your source is a 90-minute podcast, webinar, or interview, translation slots in at the end of a longer Mac pipeline: Clipolette finds the strongest moments first — pick a target length of 15, 30, 60, or 90 seconds and run Find Best Moments (⇧⌘B), with picks arriving ranked, each with a plain-language reason and a strength rating — then captions each keeper, and then you translate and export each one in the languages you need.

The AI picking runs on Apple’s on-device models and needs macOS 26 on an Apple Silicon Mac. On Macs that can’t run it — Intel machines included — scene detection still splits the source and you choose moments manually with keep toggles; the captioning and translation that follow are identical either way.

The podcast-to-shorts-on-Mac workflow covers that extraction pipeline end to end, and the YouTube-video-to-Shorts guide is the same engine pointed at existing long-form uploads. A Spanish-subtitled highlight cut from an English interview is a genuinely differentiated post — almost nobody in a given niche ships one, because until recently it took a web tool, a subtitle-file round-trip, and a budget.

Quality checklist for translated captions

Fix the source transcript first. Garbage in, mistranslated garbage out. This one step moves quality more than everything else combined.
Check proper nouns after translating. Names and brand terms occasionally get “translated.” Correct any casualties in the editable transcript and re-run.
Prefer plain phrasing when a clip is destined for translation. Idioms and wordplay in the hook are risk; plain speech crosses languages cleanly.
Watch line length in expanding languages. German and French routinely run 20–30% longer than English. The word-by-word animation handles pacing, but scan the preview for crowding and drop the caption size a notch if needed.
Mind the safe zones. Captions under TikTok’s UI stack are captions nobody reads, in any language. The Safe Zones preview shows what TikTok, Reels, and Shorts cover; the horizontal-to-vertical guide covers framing in depth.
Get a native-speaker spot check when stakes are high. For a client deliverable or a pinned post, two minutes of human review beats any engine’s confidence.

When this workflow is the right call

Your analytics show real international viewership and your clips are caption-dependent — which on muted feeds is all of them.
Your content is in a niche that travels: cooking, fitness, tech, music, gaming, crafts. Visual-first content plus translated captions is a proven pairing.
You handle footage that can’t be uploaded — NDA interviews, client work, embargoed material — which rules out every web-based subtitle translator.
You want zero marginal cost per language, per clip, at any volume.

When it isn’t

You need dubbed audio in the target language. That’s a different product category, served by subscription cloud tools; Clipolette deliberately doesn’t do it.
Your target language isn’t in Apple’s Translation framework. 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, such as YouTube long-form uploads. Clipolette burns captions into the pixels — right for short-form, wrong for that job.

The bottom line

Your clips already travel; the question is whether viewers can read them when they land. Translating video captions on Mac used to mean a web uploader, a per-minute fee, and a subtitle-file shuttle between tools. Now it’s a one-time language pack download and one more pass in the same native app that cut and captioned 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. Every tool is included free — 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 caption translation work on Intel Macs? Yes. Translation runs through Apple’s Translation framework, which is part of macOS itself, and Clipolette’s captioning and translation tools work on macOS 15.4 or later — Intel included. Only AI clip picking (Find Best Moments) requires an Apple Silicon Mac on macOS 26.

Can I export the translated subtitles as an SRT or VTT file? No. Clipolette burns the translated captions into the exported video pixels — that’s the right shape for TikTok, Reels, and Shorts, where burned-in text is the convention. If you need a subtitle file for a platform’s own caption system, that’s outside Clipolette’s scope.

Do I need macOS 26 or Apple Intelligence to translate captions? No. Captioning and translation work on macOS 15.4 and later on any Mac. Apple Intelligence on macOS 26 gates only the AI clip picker that finds the best moments in long video — on other Macs you choose scenes manually with keep toggles, and translation works the same either way.