Apple Watch Audio Intelligence enters beta with rewind and recap tools
Developers can now test features that transcribe briefly and summarise conversations, with unsaved output deleted on a short timer.
Apple has opened beta testing of its Audio Intelligence features for the Apple Watch, first shown at its September hardware event, with developers getting access through watchOS 27.3 and iOS 27.3 betas. Two of the four opt-in tools listen to conversations: Live Rewind shows the last 15 seconds of speech as a text snippet that disappears after 30 seconds unless saved, and Siri Recaps produces a short summary with bullet points that users review at day's end before saving to Notes or Journal. Apple says the recaps omit speaker identities, raw audio and verbatim transcripts, and that unsaved recaps delete after seven days. Processing runs partly on the watch's S11 chip and the paired iPhone, with the condensed text sent to Apple's Private Cloud Compute, where it is deleted right after the summary is returned. The other two features, sound recognition for alarms and doorbells and song identification through Shazam, round out the set, and Apple has also put a redesigned Health app into the iOS 27.3 beta.
Why it matters
For Apple Watch owners, the features put spoken conversation into a short-lived text form that the user chooses to keep or discard, and the summaries are built to avoid speaker names, raw audio and verbatim transcripts. Because processing is split between the watch, the paired iPhone and Apple's cloud, the privacy outcome depends on how those deletion rules are applied. The beta shows how Apple is trying to offer conversation-based assistance without keeping a full record of it.
Matters or hype?
Does this story matter, or is it hype? Vote before you see what everyone else said.