Frequently asked questions

In principle, yes for some products. An iPhone can receive Bluetooth Low Energy advertisements from nearby devices and compare them with known identifiers, and several smart-glasses makers hold registered Bluetooth identifiers. In practice, whether a specific product can be recognised, and in which states, has to be measured on real hardware. We have not completed that testing, so we do not currently claim that the app detects any specific device. See testing results and the device database.

No. Bluetooth advertisements carry no information about whether a camera or microphone is active, and no Bluetooth-based tool can determine that. The app will never display a "recording" state. The most reliable everyday cue for Meta and Snap products is the capture light the manufacturers describe; see the guide for teachers.

Unknown, and it probably varies by product and by state (in the case, in pairing mode, connected to the owner's phone, recording). Many Bluetooth accessories stop advertising once connected to their host phone. Whether the glasses we study advertise while worn and connected to someone else's phone is the first and most important question in our hardware test plan.

No. It means nothing matched in this place, over this period, with the phone in this state. Some glasses may advertise nothing while connected, iOS limits what a backgrounded app can hear, and bodies and walls attenuate signals. The app will never show an "all clear". Read false positives and false negatives.

Roughly how strong the received signal was at that moment. It is affected by transmit power, antenna orientation, bodies, walls and reflections as well as by distance, so the app shows it as a coarse category (very strong, strong, moderate, weak) and never as metres. Details on RSSI, proximity and distance.

None yet, in the verified sense. The device database lists the products we are researching together with the Bluetooth identifiers registered to their manufacturers. Every record is marked Researching or Unknown until independent hardware testing produces evidence, and the status can only change through recorded test results.

Because a company identifier is registered to a company, not a product. Meta Platforms Technologies, for example, makes Quest headsets and controllers as well as glasses. The scoring model caps any single kind of evidence below the medium threshold, so at least two independent kinds of evidence, such as a service UUID or a payload pattern, must agree. See how detection works.

Not in a way we can promise. Apple documents that background scanning requires a service-UUID filter, coalesces repeated discoveries, runs at a slower interval and can be terminated to free memory. Whether the glasses we study broadcast a usable service UUID is untested. Any background feature will be described as slower and able to miss devices, and it will only be offered if testing shows it does something useful. See what iOS lets a Bluetooth scanner do.

No. Matching happens on the phone with a bundled signature catalogue. Bluetooth observations are not uploaded, there is no account, and no location is used. Detection history stays on the device and can be deleted in one tap. The full list of promises and how each is enforced is on the privacy page.

From the Bluetooth SIG assigned-numbers registry (which company registered which identifier), from public write-ups recorded as unverified leads, and, once available, from our own hardware tests. Every signature carries its source and licence, and a community lead can never be flipped to verified in place. Nothing is copied from AGPL or non-commercial projects. Read where signatures come from.

Not yet. It is in development and has not been released on the App Store. The next milestone is physical hardware testing; the consumer app only proceeds if the results justify it. Leave your email on the notify page and we will send one message when it launches.