// ON-DEVICE AUTO-CAPTIONS / ACCESSIBILITY REVIEW

iOS 27 Generated Subtitles Review (2026): Honest Verdict on Apple's On-Device Auto-Captions — for Creators

iOS 27 Generated Subtitles review (2026): an honest creator's verdict on Apple's on-device auto-captions, its real limits, and who it actually fits best.

Last verified · 2026-09-11 · by Moe Ameen
The verdict
3.7 / 5

As an accessibility feature, Generated Subtitles is a quiet, genuinely useful win: it captions any uncaptioned video on-device, privately, systemwide, and for free, which helps deaf and hard-of-hearing viewers and anyone watching sound-off. As a creator tool it barely registers, and it isn't trying to be one — the captions are painted on the viewer's screen at playback, never written into the file, so they don't survive an export or reach anyone outside English-speaking US and Canada. Score it as what it is: a strong consumption feature and a non-starter for publishing.

Generated Subtitles is one of the accessibility features Apple previewed on May 19, 2026 and ships this fall in iOS 27, alongside iPadOS 27, macOS 27, tvOS 27, and visionOS 27. The pitch is clean: when a video has no captions of its own, the device transcribes the audio and shows subtitles automatically during playback, using an on-device speech recognition model so nothing is uploaded.

I'm reviewing it from a creator's angle, because that's who lands on a "iOS 27 captions review" search, and it's worth being upfront that this is not a product a creator buys or a step in a publishing workflow. It's a system feature aimed at the person watching. Judged on its own terms — accessibility, privacy, breadth of coverage — it's good. Judged as a way to caption content you publish, it does almost nothing, and pretending otherwise would be dishonest.

The scores below rate transcription accuracy, on-device privacy, language and region coverage, customization, systemwide reach, whether the captions travel with the file, value, and usefulness to a publishing creator. Where the feature is strong it gets credit; where it stops at the viewer's screen it's marked down for the creator use case specifically.

The one-line version: excellent that Apple is making captions a default for viewers; irrelevant to getting captioned video in front of an audience. All details reflect Apple's launch-window disclosures as of September 2026 — treat exact device requirements and rollout as a snapshot and confirm on Apple's pages.

What iOS 27 Generated Subtitles is

Generated Subtitles is a systemwide iOS 27 feature that automatically transcribes the spoken audio of any video that lacks its own caption track and displays subtitles during playback. It runs on-device using Apple's speech recognition, so the audio is not sent to the cloud, and it works across the videos you record on iPhone, videos friends and family send you, and videos streamed online — meaning it can caption a clip in Photos, Messages, a social feed, or a streaming app. Support spans iPhone, iPad, Mac, Apple TV, and Vision Pro. Viewers can adjust how the captions look from the video playback menu or in Settings, and turn the feature off. At launch it is English only and available only in the United States and Canada, with additional languages and regions expected later. What it is not: an editor, an exporter, or a caption-authoring tool. It does not produce an editable SRT, and it does not burn captions into the video file — the subtitles exist only as an overlay rendered on the watching device.

Who iOS 27 Generated Subtitles is for

Generated Subtitles is for viewers first: deaf and hard-of-hearing users, people watching muted in public, and anyone consuming videos that shipped without captions. For that audience it's an easy, free upgrade that needs no setup. For a creator, its usefulness is narrow and indirect — a private way to preview your own footage sound-off, or to check whether your audio is clear enough that a model can transcribe it. Anyone expecting it to caption the videos they publish, export a subtitle track, or style captions for TikTok or Reels is looking at the wrong tool; that's a production job it never attempts.

Scoring breakdown

DimensionScoreWhy
Transcription accuracy4.0 / 5Apple's on-device speech recognition is solid on clean English audio; muddy, accented, or overlapping speech will still trip it, as with any ASR.
Privacy & on-device processing5.0 / 5Captions are generated locally with nothing uploaded — a genuine privacy and cost advantage over cloud captioning.
Systemwide coverage4.5 / 5Works across recorded, received, and streamed video and across iPhone, iPad, Mac, Apple TV, and Vision Pro — broad by design.
Language & region coverage2.5 / 5English only, US and Canada only at launch; most of a global audience gets nothing automatic until Apple expands it.
Customization3.5 / 5You can restyle caption appearance and toggle the feature in the playback menu or Settings, but styling is for viewing, not for a branded export.
Captions travel with the file1.0 / 5They don't — subtitles are overlaid on the viewer's device and are not written into the video, so they vanish on export or when posted.
Value (free with the OS)4.5 / 5It costs nothing and needs no setup; for accessibility that's excellent value.
Usefulness for publishing creators2.0 / 5Near zero for production: no export, no burned-in captions, no per-platform styling, no distribution.

Pros and cons

Pros

  • On-device, private transcription — the audio never leaves the device, with no cloud upload or per-minute fees.
  • Systemwide: captions any uncaptioned video in Photos, Messages, social feeds, or streaming, across five Apple platforms.
  • Free and zero-setup — it just works once you're on iOS 27 in a supported region.
  • A real accessibility win for deaf and hard-of-hearing viewers and the sound-off audience.
  • Adjustable caption appearance and an easy off switch from the playback menu or Settings.
  • Useful to creators as a private, sound-off preview of their own rough footage.

Cons

  • The captions are viewer-side only — they are not burned into the file and do not travel when you export or publish.
  • No subtitle-file export (no SRT or VTT) and no editing of the transcript.
  • English only, US and Canada only at launch, so global audiences see nothing automatic.
  • No brand styling, animation, or per-platform sizing — it's a system overlay, not a caption design tool.
  • It captions nothing you distribute: no clipping, no repurposing, no scheduling, no publishing.
  • Accuracy still depends on audio quality, and there's no way to correct a mis-transcription for viewers.

Pricing analysis

There's no price to analyze in the usual sense: Generated Subtitles is bundled free with iOS 27 and its sibling operating systems. For an accessibility feature, free-and-built-in is the ideal model — it means every eligible device gets captions on uncaptioned video with no subscription, no add-on, and no per-minute transcription cost, because the work runs locally.

The catch isn't cost, it's scope. "Free" here buys a viewing aid, not a production capability. There is no paid tier that unlocks exporting a caption track, burning captions into a file, or styling them for a platform, because those aren't on the roadmap for this feature — they're simply outside what it does.

For a creator budgeting a content stack, the honest read is that Generated Subtitles is a free nicety on the consumption side and a line item you still have to fill on the production side. The transcription-for-viewing is free; captions that ship with your posts, styled and sized per platform, are a separate tool and a separate spend regardless.

Use-case fit

Use caseFitWhy
Watching uncaptioned videos sound-off or with hearing lossStrongThis is the feature's purpose — automatic, private captions on any video you watch, free and systemwide.
Previewing your own rough footage before editingOKA handy private sound-off check on the phone, though it captions for viewing only and exports nothing.
Checking whether your audio is clear enough to transcribeOKIf the on-device model struggles, that's a signal your audio needs work; but it's an indirect use.
Getting an editable SRT for a local videoWeakIt doesn't export subtitle files or let you edit the transcript at all.
Captioning a video you're about to postWeakThe captions are viewer-side and don't travel with the file, so a published clip arrives caption-less.
Styling and sizing captions per platformWeakThere is no branded styling, animation, or safe-zone sizing — it's a system overlay for viewing.
Clipping, repurposing, and publishing captioned contentWeakNone of that is in scope; it's a consumption feature, not a content workflow.

Alternatives worth considering

  • Apple SpeechAnalyzer / on-device transcription — the underlying speech tech developers can build captioning on, if you need the transcript itself rather than a viewing overlay.
  • Final Cut Pro Generate Captions — Apple's in-editor auto-captioning that actually places and styles subtitles in the timeline for export, for people editing on Mac or iPad.
  • Dedicated caption apps (e.g. Submagic, CapCut auto-captions) — burn styled captions into short-form video for posting, which Generated Subtitles never does.
  • Kompozy — not a captioning feature but a content generation and publishing engine that transcribes, clips, burns in per-platform styled captions, and distributes across platforms.

How Kompozy compares

To be straight, since I run Kompozy: it is not a competitor to Generated Subtitles and the two shouldn't be compared head-to-head. Generated Subtitles captions video for the person watching; Kompozy captions video for the world. The only reason it belongs in this review is the creator who arrives assuming Apple's auto-captions will handle the videos they publish. They won't — the subtitles live on the viewer's screen and never touch the file.

Kompozy is where that gap gets filled. It takes a raw recording, transcribes it, cuts it into vertical Clipped Shorts, and burns word-synced captions into each cut with a chosen style and per-platform placement — captions written into the pixels, seen by every viewer on any device in any region. From the same source it also builds carousels, quote graphics, a blog, and a newsletter under one Persona Brief, then schedules and publishes across the eight social platforms plus blog and email. If you're watching video, Generated Subtitles is a free, welcome upgrade. If you're publishing it, that's Kompozy's job, and the handoff is clean.

Frequently asked questions

Is iOS 27 Generated Subtitles good?

As an accessibility feature, yes — it captions any uncaptioned video on-device, privately, systemwide, and free, which is a real win for deaf and hard-of-hearing viewers and anyone watching sound-off. As a creator tool it does very little, because the captions are viewer-side and never written into the file.

Does it work offline and keep my videos private?

Yes. Generated Subtitles uses an on-device speech recognition model, so the audio is not uploaded to the cloud and captions are generated privately. That on-device approach is one of its clearest strengths.

Can I export the captions or an SRT file from it?

No. It is a viewer-side feature that overlays subtitles during playback. It does not export an editable SRT or VTT file and does not let you edit the transcript. For an exportable caption track you need a transcription or editing tool.

Do the captions stay on a video when I post it?

No. The subtitles are rendered on the watching device and are not burned into the file, so when you export or publish the clip they disappear. To caption a published video for every viewer, the captions have to be baked in — an engine like Kompozy does that and sizes them per platform.

What languages and regions does it support at launch?

English only, and only in the United States and Canada at launch. Apple is expected to add more languages and regions later. Until then, content for a global audience still needs captions baked into the video itself.

Which devices support Generated Subtitles?

Apple lists support across iPhone, iPad, Mac, Apple TV, and Vision Pro — iOS 27, iPadOS 27, macOS 27, tvOS 27, and visionOS 27. Treat specific hardware requirements as a launch-window snapshot and confirm on Apple's pages.

Is it a good tool for captioning content I create?

Not really. It captions video for viewing, not for publishing — no export, no burned-in captions, no branded styling, no per-platform sizing, and no distribution. For creating captioned content, use an editor like Final Cut Pro or a generation-and-publishing engine like Kompozy.

Related deep guides

See iOS 27 Generated Subtitles vs Kompozy comparison → · Get Started →