Meta AI glasses sales boom as detection apps emerge
As Meta rapidly scales production of its AI smart glasses to tens of millions of units, independent developers are launching Bluetooth detection apps to combat nonconsensual recording.
Meta is aggressively expanding its hardware footprint, having sold more than 7 million AI glasses in 2025—a massive leap from the 2 million units sold over the previous two years combined. The company plans to double its Ray-Ban output, aiming to produce between 20 million and 30 million smart glasses annually by the end of 2026. This rapid proliferation has triggered intense privacy concerns, prompting venues like DEF CON 2026 to ban the devices entirely. In response, independent software developers are building Bluetooth-based detection apps to help the public identify when these devices are nearby.
Several consumer-facing detector apps have emerged to scan for the Bluetooth signals emitted by Ray-Ban Meta, Oakley Meta, and Snap Spectacles. The most popular app, Nearby Glasses, developed by Yves Jeanrenaud, costs $2.99 on iOS and Android and has surpassed 100,000 downloads on the Google Play Store. Another iOS app, Antizuck Smart Glasses Scanner, also sells for $3. A newer free iPhone app called Zuckoff, created by Pawel Szydlowski, offers an optional background-scanning subscription for $10 a year and is currently seeking 12 testers to launch on Android. These apps estimate distance based on signal strength to flag whether a device is within arm's length, in the same room, near, or far.
For software developers and privacy practitioners, these apps highlight the technical challenges of tracking wearable surveillance. The detectors rely on Bluetooth Low Energy signals and specific regulatory product identifiers, meaning they cannot confirm if a camera is actively recording. This limitation is compounded by Meta's hardware design; the physical recording LED can easily be covered by a sticker. Furthermore, Meta is reportedly testing a super sensing prototype that may disable the LED entirely during continuous recording, and the company recently deleted an unreleased facial recognition system that had been quietly installed on 50 million phones in June. For developers in the privacy space, this cat-and-mouse dynamic underscores the need for standardized broadcasting protocols, such as legally mandated Bluetooth identifiers, to make wearable AI devices transparent to the environments they record.
This is our own summary of reporting by Ars Technica AI
