Frequently Asked Questions
Answers about compatible signal recognition, product limits, and what happens to your data.
What NearLens does, and what it does not
Product transparency
Coverage and reliability
NearLens recognises only Bluetooth identifiers and observable signal patterns that its current recognition data can evaluate. It is not a universal detector, and no result is proof that smart glasses are absent.
- Recognition coverage
- A confirmed result depends on a registered identifier or verified signal signature already present in NearLens recognition data. New, generic, or undocumented devices may not produce a confirmed match.
- Known sources of false positives
- Headsets, earbuds, speakers, and other products from the same manufacturers may expose related Bluetooth information. NearLens reports signal compatibility, not visual proof of a specific device.
- Known reasons for missed detections
- A device cannot be detected when it is not broadcasting, Bluetooth is off, it is asleep, its advertisement is suppressed, or its identifiers and signal pattern are not yet recognised.
- Distance and testing
- Distance is estimated from signal strength and can change with walls, crowds, interference, and device behaviour. Field observations are not a controlled accuracy benchmark, so NearLens does not publish an unsupported accuracy rate.
What confidence means
A confidence label is not a recording probability. It describes how strongly the observable Bluetooth information matches recognition data. At every level, NearLens reports a nearby signal, not who is carrying a device or what that device is doing.
- High confidence
- The strongest match, based on a registered identifier or verified signal signature associated with a recognised smart-glasses profile.
- Medium confidence
- Several observable signal details are consistent with a smart-glasses profile, but the information is not strong enough for the highest-confidence match.
- Low confidence
- A partial or ambiguous signal pattern may be compatible, but an unrelated device is more plausible. Treat it as a weak indication and use your judgement.
The basics
NearLens scans for Bluetooth signals from smart glasses and tells you when one is nearby. You get a discreet alert with an estimated distance and the level of confidence behind the match.
The purpose is awareness. Knowing a camera may be in the room lets you decide how to act, which is a choice you do not currently have.
No. NearLens is not about the people who wear smart glasses, and most of them are not doing anything wrong. Many use them for music, calls, navigation, translation, or accessibility.
NearLens reads a Bluetooth signal, never a person. It does not identify anyone, record anything, or tell anyone what to do. The Terms of Use prohibit using NearLens to approach, confront, or target anyone.
Some brands make models with and without cameras, and a few share the same Bluetooth identifiers across products. When NearLens cannot tell models apart, the result is shown at lower confidence.
If camera-free glasses are appearing, email hello@marcosrezende.com with the model name so the recognition data can be improved.
NearLens detects devices whose Bluetooth identifiers or verified signal signatures are already in its recognition data. It is not a universal detector.
Some low-cost or generic frames do not broadcast a detectable Bluetooth signal, while others use identifiers or payloads that are not publicly documented. Those devices may be nearby without producing a result.
No. NearLens detects that a device is present. Recording state is not broadcast over Bluetooth, so no app can read it from outside the glasses.
Presence is still useful information. A camera that might be recording is a different situation from no camera at all.
No, and it never will. NearLens reads a device signal, not a person. There is no identity, no photo, no name, and no way to link a detection to an individual.
There is also no map, no shared feed, and no way to send a detection to another user. Those features were left out on purpose, because a tool built to protect people should not make it easier to target them.
Most people wearing smart glasses are not recording anyone. But when a camera is on a face, people nearby have no easy way to know it is there.
NearLens gives them that information, nothing more.
Knowing that a compatible device may be nearby lets you decide what feels right for you, like staying, moving, or asking venue staff about their policy.
The app returns one signal: not proof of recording, not the wearer's identity, just presence.
The honest position is that NearLens is useful and insufficient. The structural fix is consent design at the platform level, not a detection app. NearLens treats a symptom, and it does not pretend otherwise.
Yes, by design. A phone camera requires the user to raise a device to their face, which is a visible act. Smart glasses have no equivalent signal: the camera sits inside the frame, there is no viewfinder, no shutter sound unless the manufacturer adds one, and no indicator that bystanders can read from a normal distance.
Ray-Ban Meta, for example, has a small white LED that lights while recording, but it is easy to miss in ambient light, and at least one documented modification removed it entirely. NearLens exists because the visual cue that phone cameras provide does not reliably transfer to wearables.
NearLens recognition data includes observable identifiers or verified signal signatures associated with Ray-Ban Meta, Oakley Meta, Snap Spectacles, and compatible EssilorLuxottica smart eyewear.
Coverage changes as manufacturers release new products or update Bluetooth behaviour. NearLens can only recognise devices whose observable identifiers or verified signal signatures are in its recognition data.
It is a fair objection. Passive Bluetooth scanning reveals that devices are present nearby, and in a small enough space that signal can narrow things down without a name attached.
NearLens does not solve privacy. It is one initiative in a structural problem that requires a platform-level response: consent design built into the hardware, regulatory pressure on manufacturers, not a detection app. The app makes no claim otherwise.
What it contributes is a single signal: presence. The wearer of smart glasses already knows exactly where the camera is pointed. NearLens gives the person on the other side a rough equivalent. No names, no faces, no logs transmitted anywhere. That is enough to make a decision where none was previously possible. That is the extent of what it sets out to do.
Your data
The NearLens app does not use analytics, telemetry, advertising, or account data. Bluetooth observations and detection history stay on the device.
The website is separate and uses optional Umami Cloud analytics only after consent. It also uses one necessary cookie to remember that choice.
NearLens requests Bluetooth access to scan for nearby BLE signals. iOS does not require location authorisation for Core Bluetooth scanning.
NearLens does not need GPS coordinates for detection.
Detection history is opt-in and stored locally. NearLens does not publish a fixed automatic-clearing interval.
You can clear the full log at any time from within the app, and deleting the app deletes its local history.
Only if you turn it on. Save Detection History is off by default and lives entirely on your device.
You can clear it at any time, and deleting the app deletes the history with it. Nothing is backed up to a server, because there is no server.
The NearLens app processes Bluetooth observations, detection history, and Privacy Intelligence on the device. It does not send that information to this website.
The website uses the internet to serve pages, optional consent-controlled analytics, and optional press translation.
No. NearLens only listens for the advertising packet that every Bluetooth device broadcasts openly to announce itself. No pairing is attempted, nothing is intercepted, and nothing is decrypted.
The wearer is not contacted, notified, or affected in any way.
Detection and accuracy
NearLens listens for Bluetooth Low Energy advertising packets that a device makes visible nearby.
It compares observable identifiers and verified signal fields against recognition data for known smart-glasses hardware. It cannot identify a device that is not advertising a detectable signal or whose identifier or signature is not in that data.
Bluetooth Low Energy devices broadcast short-range advertising signals. NearLens listens for compatible signals associated with supported smart-glasses hardware.
That provides presence awareness, not person tracking, a map, or proof that a device is recording. A missing result only means that no compatible signal was observed.
Bluetooth devices can rotate the address they use in advertising packets. The same pair of glasses can therefore appear as several advertising identities over time, while two entries cannot safely be assumed to be two separate devices.
NearLens reports recognised signals for awareness, not a count of distinct glasses or people.
No. NearLens can recognise a compatible device signal, but it cannot read the device's firmware version, current connection state, recording state, or listening state from outside the glasses.
A model match describes the hardware, not everything that unit is doing now.
No. NearLens does not see where another device sends captured images or audio.
A Bluetooth advertising signal does not reveal the cloud provider, processing location, or retention policy chosen by the wearer. Those details require the manufacturer's documentation or independent testing, not a nearby-device scan.
A confirmed match means that the observed Bluetooth information matched a registered identifier or verified device signature. It does not mean that every pair of glasses in the room was found.
If a device broadcasts incompletely or uses an unpublished identifier, NearLens may not be able to make a reliable match.
NearLens needs two things: a detectable Bluetooth signal and an identifier or verified signal signature that the app recognises.
Some low-cost or generic frames do not transmit Bluetooth at all, and some devices use identifiers or payloads that are not publicly documented. No result means that no compatible signal was observed. It is not proof that no camera-equipped glasses are nearby.
No. A device that is not broadcasting cannot be detected by any Bluetooth scanner. NearLens sees no more than the glasses choose to announce.
The same applies to walls and direction. Signal strength gives a rough distance, not a location.
NearLens has been used in field observations, but those observations are not a controlled accuracy benchmark.
Detection depends on a compatible Bluetooth broadcast, recognition data, distance, interference, and device behaviour. No result is proof that no smart glasses are present, and a result is not proof of recording.
Bluetooth Low Energy range depends on the environment. In open space, NearLens can pick up advertising packets from 10 to 15 metres away. In a dense venue with many competing signals and physical obstacles, that drops to roughly 5 to 8 metres.
The app labels each detection with an estimated distance (within arm's reach, same room, or nearby), derived from signal strength, not precise measurement. Range is also affected by whether the glasses are facing toward your phone and whether the wearer has them paired to another device.
Sometimes, but only when the device broadcasts a detectable Bluetooth signal and its identifier or verified signal signature is in NearLens recognition data.
Many low-cost or generic frames do not transmit Bluetooth, and some use identifiers or payloads that are not publicly documented. In those cases, the device may not be detected even if it is nearby.
Using the app
Privacy Intelligence provides local summaries of detection patterns, including repeat encounters and cluster events, with an exportable report.
Core Bluetooth detection does not depend on a generative model. A natural-language Q&A using Apple Foundation Models is planned.
The Privacy Intelligence report summarises local detection history, including frequently observed brands, active time periods, repeat encounters, and cluster events.
The report can be exported. A natural-language Q&A is planned.
Yes. Local Privacy Intelligence summaries are available in the current NearLens release. The base app can be installed on iOS 15.1.
A natural-language Q&A using Apple Foundation Models remains planned.
Yes. Snooze for 15 minutes, for one hour, or for the rest of the session.
Alerts are haptic first and use pink rather than red, so a detection stays private to you and reads as information instead of an emergency. Constant agitation is not the goal.
A detection is not an accusation. It means a compatible signal is nearby, not that anyone is recording or doing anything wrong.
An alert is not an instruction to run or confront anyone. NearLens observed a compatible Bluetooth signal nearby.
Pause and choose what makes sense: move to another seat or room, avoid a sensitive conversation, ask venue staff about their recording policy, or carry on. If someone is making you feel unsafe, prioritise your safety and seek help from staff or local emergency services.
NearLens cannot tell who is wearing the device, whether it is recording, or what they intend. Treat the alert as information for your decision, not proof that anyone did something wrong. Using NearLens to accuse a stranger would recreate the problem the app was built to address.
NearLens is free. There is no subscription, no advertising, and no data to sell because none is collected.
Android is not currently supported. The current NearLens experience is designed for iPhone.
NearLens focuses on recognising Bluetooth signals associated with camera-capable smart glasses. General-purpose Bluetooth scanners list every device in range without smart-glasses classification or confidence labels.
Other smart-glasses detector apps also exist. NearLens should be compared by its documented capabilities and limits, not by an exclusivity claim.
No. There is no paid tier and no features held back. Every user gets the same app.
Being free does not change what the app can and cannot detect.
