watchOS 27.3 Beta Adds Siri Recap and Live Rewind to Apple Watch

The watchOS 27.3 beta now includes Siri Recap and Live Rewind, two Audio Intelligence features for Apple Watch, according to Engadget.

Apple first disclosed both features in September, when it unveiled the Apple Watch Series 12 and Apple Watch Ultra 4. In its Siri announcement, Apple said Live Rewind and Siri Recap will be available in beta in late 2026. Both features run on Series 12 and Ultra 4.

Siri Recap produces high-level summaries of conversations. Each summary carries a title and key points generated by Apple Intelligence, Apple's AI system, and summaries are accessed from the Siri app. Apple says Recap summaries will not surface sensitive information such as financial details.

Live Rewind works on a much shorter window. It converts the last 15 seconds of a conversation to text, a rolling buffer that always holds only the most recent stretch of speech. Double-pressing the Digital Crown, the rotating side button on the watch, displays the resulting snippet. Users can query that text through Siri, or save it to the Siri app to read later.

Both features depend on Apple Intelligence. They require a compatible iPhone with Apple Intelligence enabled, which Engadget defines as an iPhone 16 or later, excluding the iPhone 16e. Siri Recap and Live Rewind are initially limited to English and are not available in the EU. Both are also subject to daily usage limits.

On privacy, Apple says Audio Intelligence features do not record or store audio. According to Apple, the S11 chip processes audio in isolation and deletes it immediately. Apple states that its operating systems, apps, users, and Apple itself cannot access the raw audio used for these features. Apple also says Audio Intelligence does not identify or attribute speakers in a conversation.

Live Rewind makes its activity visible to people nearby. It plays a chime when in use, even if the Apple Watch is set to silent, and the display shows an animation and a microphone indicator.

In my view, the privacy design choices will be the part of this release that technical readers examine most closely. The audible chime and on-screen microphone indicator are the signals a bystander can actually perceive, and they give the feature a visible footprint that ambient transcription often lacks.

The broader context here is the shift of on-device and private-cloud language models from the phone to the wrist. A 15-second rolling transcript is a narrow capability, but one that can recover a missed phrase, an address, or a name without the user having started a recording beforehand. Whether the daily limits, the English-only launch, and the iPhone 16 requirement shape real adoption will become clearer as the beta matures toward the late-2026 timeframe Apple has described.