One earbud takes one photo, the two are then merged into one—this isn't some concept design, but rather the actual operating logic that's now being pieced together after Apple's unannounced camera AirPods were accidentally leaked.
Earlier this week, Apple accidentally leaked a demo video in the macOS Tahoe 26.7 beta, exposing this long-rumored earbud. Now more details have surfaced buried in the code of that same early build: each earbud is equipped with its own independent camera, and the footage captured is synced and merged into a single complete image by Visual Intelligence software. The camera specs aren't impressive—capable of shooting RGB still photos at a maximum resolution of 1 megapixel.
Given that the head is constantly moving, the system has a corresponding mechanism built in: AirPods already come with multiple built-in accelerometers, and the code shows the device detects and discards frames that are distorted due to sudden movement or blocked viewing angles. Some processing appears to happen directly on the earbuds themselves—the code contains the term "peripheral inference," suggesting this is a form of real-time on-device interpretation rather than something handed off entirely to the phone or cloud.
It's worth clarifying that these AirPods don't record video—they only take "periodic photos," and Apple apparently believes this amount of information is enough for Visual Intelligence to understand the scene in front of the wearer. The device has both active and passive modes, with higher image quality in active mode; the code also indicates each earbud may be equipped with a capture indicator light to signal that the camera is in operation.
