Meta’s AI Glasses Can Now Hear Conversations Better Than You

Meta has announced a batch of new audio features tied to its latest software update for its AI-powered smart glasses, including improvements to sound processing, music controls, and fitness features.

The update includes the release of SAM Audio, a new AI model designed to separate sounds from audio or video clips. Meta says the tool lets users isolate specific sounds using text prompts, visual cues, or by selecting a time segment in the audio.

The feature is part of Meta’s broader “Segment Anything” AI project and is aimed at making basic audio editing easier for creators without professional tools.

Alongside SAM Audio, Meta is also rolling out updates to how audio works on its Ray-Ban Meta and Oakley Meta smart glasses.

One of the biggest changes is a new “Conversation Focus” feature, which is designed to make voices easier to hear in noisy environments. Meta says this feature is entering early access in the U.S. and Canada for Ray-Ban Meta and Oakley Meta HSTN glasses.

Meta is also expanding music features. Users can now ask Meta AI to play music based on what they are looking at, such as saying “Hey Meta, play a song to match this view.” The feature uses the glasses’ cameras and AI to create a playlist based on the scene and the user’s listening history. Music voice controls are also expanding to support French, German, Italian, Portuguese, and Spanish.

For fitness-focused users, Meta is adding new hands-free workout features for the Oakley Meta Vanguard glasses. When paired with a compatible Garmin device, users can create custom running or cycling workouts directly from the glasses. Meta is also adding new single-word voice commands like “photo,” “video,” and “music” to make controls quicker during workouts.

The updates are part of software version 21 for Meta AI glasses and will roll out gradually, with availability depending on device model and region.