Smart Glasses Announce Themselves Over Bluetooth, and a $2.99 iPhone App Now Listens for It

Every wireless accessory you own announces itself. That is not a design flaw, it is how Bluetooth Low Energy works: a device that wants to be findable broadcasts short advertising packets into the air on a repeating interval, and any radio in range can read them without pairing, without permission, and without the broadcasting device ever knowing it happened. Earbuds do it. Fitness trackers do it. Smart glasses, which are architecturally much closer to earbuds than to phones, inherited the behavior along with the rest of the accessory stack.

Somebody finally built the obvious app.

AntiZuck went up on the App Store as a one-time $2.99 purchase for iPhone and iPad, and it does one thing: it listens for Bluetooth advertisements whose signatures match known smart glasses and tells you when it thinks a pair is nearby. The supported list covers Meta Ray-Ban glasses, Meta's own-brand frames, Snap Spectacles, Amazon Echo Frames, and RayNeo. Detections surface as blips with an estimated proximity and a confidence level, plus a history of repeat sightings, optional local alerts that fire when the app is closed, Home Screen and Lock Screen widgets, and a Control Center toggle that can be mapped to the Action Button.

The restraint in the design is what makes it worth a developer's attention. AntiZuck does not attempt to identify the person wearing the glasses, does not claim to know whether a device is recording, and does not report direction. Everything happens on device. No account, no analytics, no uploaded signal history, no location access, no camera or microphone permission. An app built to make people less comfortable being surveilled would have been trivially easy to build as a surveillance product itself, and this one conspicuously is not.

The detection window is narrower than the marketing suggests

Here is the part that a developer will notice before anyone else does. Smart glasses do not broadcast continuously, because a radio that advertises nonstop is a radio that eats a battery measured in single-digit hours. They advertise in bursts, during the events that require discoverability: powering on, entering pairing mode, or the charging case being opened. A pair that has been sitting on someone's face, already bonded to their phone, quietly doing nothing, may not produce a single packet the whole time it is near you.

AntiZuck's developer says as much, describing detection as best effort. That honesty is correct and it is also the whole ballgame, because it means the tool is far better at catching the moment someone takes their glasses out of the case in a coffee shop than at telling you whether the person across the train car has been wearing a camera for twenty minutes. Useful. Not a radar dome.

AntiZuck is not alone, either. Nearby Glasses shipped earlier this year working from the same primitive, cross-referencing observed advertisements against a database of known device signatures, and it got enough attention to prove there was demand for the category well before this month's round of coverage. Two independent implementations of the same idea, built by unaffiliated developers, is usually the point at which a hobby project becomes a category.

The allowlist feature is the tell

AntiZuck lets an owner register their own glasses so their personal hardware stops setting off their own alerts, moving those devices into a separate list where the last detection time is still visible. Read that as a product decision and it is thoughtful. Read it as a technical claim and it is more interesting than that, because allowlisting a specific pair requires the observed signature to be stable enough to recognize the same physical device across sessions.

Which raises the question every platform team at Meta, Snap, and Google is presumably now discussing internally. The BLE specification has supported resolvable private addresses for years precisely so that a device cannot be tracked by its MAC, and most modern accessories rotate them. If glasses remain reliably fingerprintable anyway, the identifier is living somewhere else, most likely in manufacturer-specific data or in the set of service UUIDs a device advertises, both of which are far easier to obfuscate than most people assume.

So the vendors can make this harder whenever they decide to. That is the uncomfortable part. Randomizing the advertised payload would break every detector app on the market inside one firmware update, and it would also be a company choosing, on the record, to make its camera-equipped product harder for bystanders to notice. Meta has already been reported to be reconsidering the physical privacy LED on its next generation, which was the previous best answer to this problem and which only ever worked if someone happened to be looking directly at your face. Bluetooth detection works at range, through a jacket pocket, with nobody making eye contact at all.

This eventually stops being a third-party app

Apple already ships unwanted-tracker detection at the operating system level, because AirTags forced the issue and the industry landed on a cross-platform specification rather than leaving it to the App Store. The underlying capability is identical. A phone that can tell you an unknown tracker has been moving with you can tell you an unknown camera has been sitting next to you, and the only real difference is which advertising signatures the OS chooses to care about.

Whether Apple or Google wants that fight is a separate matter, and both have obvious reasons to hesitate, given that Google is currently shipping an Android XR glasses platform with Samsung and Apple has its own frames slated for a WWDC 2027 unveil. Neither company is eager to build the feature that makes its own future product look like a threat.

For now the capability sits with independent developers charging three dollars, working from a broadcast the hardware makers never intended as a public identifier. That is roughly how the tracker-detection category started too.