What happened
An open-source project offers Android and iOS apps that attempt to detect nearby smart glasses through Bluetooth Low Energy advertising data and warn users. The repository documents the app's detection methods, distribution options and significant technical limitations.
The GitHub repository for Nearby Glasses describes an open-source app whose sole stated purpose is to look for nearby smart glasses and warn the user. It provides Android and iOS source code under the AGPL-3.0 license. The README says Android users can install builds from project releases or Google Play, while an iOS version is available through the Apple App Store and as a sideloadable package. The repository dates the iOS version to March 24, 2026. Those distribution and availability statements come from the repository; the supplied source does not independently verify the current store listings or the app's reach.
The app does not claim to identify a person or prove that a recording is taking place. Instead, it looks for signals associated with manufacturers and products. The README says randomized Bluetooth addresses and unstable service identifiers make simple beacon tracking unreliable, so the app examines manufacturer identification codes in Bluetooth advertising packets, broadcast device names and selected service UUIDs. Its built-in list includes identifiers associated with Meta, Luxottica, Snapchat and another electronics manufacturer. Users can also supply their own company IDs. This makes the system a heuristic matching tool rather than a confirmed smart-glasses detector.
The project uses a default received-signal-strength threshold of -75 dBm to decide when a matching device is close enough to trigger an alert. The README gives rough open-space and indoor distance estimates, while also noting that walls, bodies, antenna design, reflections, interference and device orientation affect the measurement. It does not provide a controlled accuracy study, a tested false-positive rate or a reproducible field-test methodology. The author explicitly says the app may mistake virtual-reality headsets and other products for smart glasses, and may also miss nearby glasses because Bluetooth Low Energy detection does not always work as expected.
Read the primary source: github.com ↗
Why it matters
The project represents a practical privacy-defense attempt as camera-equipped glasses become a public concern, but a Bluetooth alert does not establish that a person is recording or even wearing smart glasses. Misidentification could create privacy risks of its own.
The project turns an abstract privacy concern into a tool that people can run on a phone. If a user is worried about camera-equipped glasses in a crowded place, a warning may prompt them to think about their surroundings or seek information about local privacy rules. But the warning is only evidence that a compatible Bluetooth signal was observed. It cannot establish which person owns the device, whether the device has a camera, whether its camera is active or whether any recording is happening.
That distinction is a safety issue. The README itself warns that harassing someone because a device appears nearby could be unlawful and tells users not to act rashly. A Bluetooth signal can travel through a room, reflect from surfaces and come from a device that is not visible. A false alert could therefore direct suspicion toward an innocent person, while a missed alert could create unwarranted confidence that no smart glasses are present. The source gives no evidence that the app can reliably connect a signal to a specific wearer or location.
The repository also makes broader claims about smart-glasses surveillance, including assertions that recording indicators can be disabled and that facial recognition has been used or may become available. In the supplied source, those statements are presented by the project author rather than established through reported tests, regulatory findings or manufacturer documentation. The repository likewise does not establish that Meta is introducing a new capability or that smart-glasses surveillance is increasing. The directly supported development is narrower: an individual open-source project is attempting to detect some devices and distribute warnings.
The app's own privacy design is relevant but also limited. The README says the app has no telemetry, advertising or automatic data collection, and that logs remain local unless users export them. It also acknowledges that exported logs contain manufacturer identification codes and could carry sensitive information depending on where they are shared. The source does not include an independent privacy audit. Nor does it establish what information Google or Apple may collect through their app-distribution systems beyond the author's general warning that platform providers may collect statistics.
What to watch next
Independent testing would be needed to measure detection and false-alert rates across devices and environments. Future changes to Bluetooth identifiers, operating-system restrictions, app distribution and safeguards against harassment will determine whether the tool is useful or misleading.
The most important missing evidence is independent, repeatable testing. Useful evaluations would compare the app's alerts with known smart-glasses models, virtual-reality headsets and unrelated Bluetooth devices across indoor and outdoor settings. They would also test crowded environments, different phone models, battery-saving settings, walls and changing distances. The repository's own warnings mean that the number of correct detections, false positives and missed devices matters more than the existence of a notification feature.
Identifier stability is another key issue. The README says some manufacturers use proprietary Bluetooth services and that service UUIDs may not remain persistent. It also says the project may add more identifiers over time. Changes in advertising behavior, operating-system permissions or product firmware could make existing rules less useful, while broad manufacturer-level identifiers could continue to match unrelated equipment. Future releases should therefore be assessed for documented changes, device-specific testing and whether new detection rules improve precision without reducing coverage.
The project's platform split also deserves attention. Android can use a foreground service to keep scanning, according to the README, while iOS cannot perform the same continuous background scan under the author's current implementation and instead offers a canary mode. That difference may materially affect what users believe the app can detect. Store availability, permission requirements and battery impact should be verified for each platform rather than generalized from one version.
Finally, any wider adoption should be considered alongside safeguards against misuse. The source provides settings for notification cooldowns, signal thresholds, logging and custom company IDs, but it does not report abuse-prevention features, harassment incidents or legal outcomes. A responsible assessment would examine how warnings are presented, whether users understand their uncertainty, and how exported logs are protected. The repository offers a technical experiment and a privacy-oriented use case; it does not yet demonstrate a dependable way to identify a nearby recording device or its wearer.
