Apple's systemwide, on-device auto-caption feature — it transcribes any uncaptioned video during playback so you can watch sound-off. English in the US and Canada at launch.
Last verified · 2026-09-11 · by Moe Ameen
Generated Subtitles is an accessibility feature Apple previewed on May 19, 2026 and ships this fall in iOS 27 (and iPadOS 27, macOS 27, tvOS 27, and visionOS 27). When a video has no caption track of its own, the operating system transcribes the spoken audio and displays subtitles automatically during playback — no button, no upload, no separate app. Apple's framing: videos "can display transcriptions of spoken audio automatically when captions or subtitles are not already provided."
The defining detail is that it runs entirely on-device. Apple uses an on-device speech recognition model so the subtitles are generated privately and the audio is never sent to the cloud. It works systemwide — across videos you record on iPhone, videos friends and family send you, and videos streamed online — so it can caption a clip in Photos, in Messages, in a social feed, or in a streaming player. Viewers control how the captions look from the video playback menu or Settings and can switch the feature off.
At launch it is English only and limited to the United States and Canada, with more languages and regions expected to follow Apple's usual rollout cadence. It is worth being precise about what this is: a consumption feature. It captions video for the person watching, on their device, in the moment. It is not an editor, it does not export a subtitle file, and it does not write captions into the video. Treat device requirements and rollout timing as a launch-window snapshot and confirm details on Apple's own pages.
iOS 27 makes captions a solved problem for your eyes; it does nothing for your posts. Because the subtitles are painted onto the viewer's screen at playback and gated to English in two countries, they exist only while an iOS 27 user watches — the moment you export the clip and post it, the captions are gone. Publishing needs the opposite: captions rendered into the pixels of the file, styled to your brand, and sized to each platform's safe zones so every viewer sees them regardless of device, OS, or region. That production step is where [Kompozy](/) picks up.
Kompozy runs the whole caption-to-publish pipeline that Generated Subtitles has no concept of. Drop a long recording in and it transcribes the audio, cuts the take into vertical [Clipped Shorts](/glossary/persona-shorts), and burns word-synced [captions](/glossary/caption) into each cut with a chosen style, animation, and per-platform placement — permanent, not overlaid. The same source also fans into carousels, quote graphics, Photo Posts, a blog article, and a newsletter, every asset held to one [Persona Brief](/glossary/persona-brief) so the channel reads as one voice, and [Autopilot](/glossary/autopilot) schedules and publishes the set across the eight social platforms plus blog and email behind a per-post review. Use iOS 27 to preview footage sound-off; use Kompozy to turn that footage into captioned, on-brand posts that ship everywhere.
No. Generated Subtitles are drawn on the viewer's device during playback and are not written into the video file. When you export or publish a clip, the captions do not travel with it. To caption a published video for every viewer, the captions have to be burned in — a content engine like Kompozy does that and sizes them per platform.
No. It is a viewer-side accessibility feature, not an editor. It transcribes audio for on-screen display and does not export an editable SRT or VTT subtitle file. If you need a caption track to work with, that comes from a transcription or editing tool, not from Generated Subtitles.
At launch it supports English only and is available in the United States and Canada. Apple is expected to add more languages and regions later. For content aimed at a global audience, captions still have to be baked into the video itself.
iOS 27 captions video for the person watching, on their device, in the moment. Kompozy captions video for the world: it transcribes, clips, and burns styled, per-platform captions into the file, then publishes across nine platforms. One is for consumption; the other is for production and distribution.