Disclosure: Appsera is building a GPS and tracker detection app for iPhone, so we have a commercial interest in this topic. That is exactly why this page leads with the limits rather than the features. Every capability claim below is tied to a sensor, an API or a published study, and where a technique does not work we say so instead of selling around it.

The four kinds of vehicle tracker

Almost every confused article on this subject fails because it treats “GPS tracker” as one thing. It is four things, they work in completely different ways, and the detection method that finds one is useless against the others. Sort out which one you are worried about and the rest of this page becomes simple.

Two practical consequences fall straight out of that table. First, if you are worried about an AirTag, your phone is already better equipped than any app you could download, and the next two sections explain why. Second, if you are worried about a professional-grade cellular tracker or a passive logger, no phone app will ever find it, and the twenty minutes you spend physically searching the car is worth more than every detector app on the store combined.

What a phone genuinely detects, sensor by sensor

A phone is a bundle of specific sensors, each with a specific job. Detection claims are only credible if you can name the sensor doing the work. Here is the honest inventory.

Bluetooth LE scanning: real, and the strongest tool you have

Every phone can listen for Bluetooth Low Energy advertisements, and consumer tracking tags advertise continuously by design, because that is how they get found. This is the one detection route that genuinely works on a phone, and it works because the tag is cooperating.

The research foundation here is solid. Heinrich, Stute, Kornhuber and Hollick reverse-engineered Apple’s offline finding protocol and published it in Proceedings on Privacy Enhancing Technologies in 2021, measuring roughly 10 metre location accuracy in urban areas. The same group then built AirGuard, presented at ACM WiSec in 2022, on the explicit premise that Apple’s own Tracker Detect app for Android gave Android users insufficient protection.

There is a catch that almost nobody mentions, and it matters if you are choosing between an app and the operating system. On iOS, an app never sees a Bluetooth device’s MAC address. Apple’s Core Bluetooth framework hands apps an operating-system-generated UUID instead, and that UUID is not guaranteed to stay the same between scanning sessions. Tracking tags also rotate their advertised identifier on a schedule. Put those two facts together and a third-party iOS app cannot reliably say “this same device has been near you for the last hour” using identifiers alone. The operating system can, because it has hardware-level access that no app is given. That is the single clearest reason to trust your phone’s built-in alerts over a downloaded scanner.

The magnetometer: real, but it is a contact-range tool

Your phone’s compass is a three-axis magnetic sensor. A common part used in phones resolves around 0.15 microtesla per step across a range of roughly plus or minus 4,900 microtesla. Earth’s magnetic field is about 25 to 65 microtesla for comparison. A magnetic-mount tracker case contains neodymium magnets, and those magnets are detectable.

The problem is geometry, not sensitivity. A magnetic field falls off with the cube of distance. Double the distance and the field drops by a factor of eight. That single fact is why magnetometer apps disappoint people: a magnet that reads clearly when the phone is touching it can be below the sensor’s noise floor a metre away. Magnet manufacturers who have tested this on iPhones report exactly that, and also warn that a strong magnet held against a phone can permanently corrupt the magnetometer’s calibration.

So the honest description of a magnetometer app is: a contact-proximity metal finder. Swept slowly, almost touching the surface, along a wheel arch or a bumper, it can flag the steel and magnets in a tracker housing. Waved around the inside of a car from the driver’s seat, it tells you nothing.

The things that will set it off anyway

• The car itself. Chassis members, seat frames, wheel arches and the exhaust distort the field far more than a small magnet does.
• Speakers. The magnets in door and parcel-shelf speakers are much stronger than a tracker mount.
• Motors and wiring. Window motors, seat motors, the blower, the alternator, and direct current running through the loom.
• Your own phone case. A MagSafe case, a magnetic wallet case or a magnetic car-mount plate sits millimetres from the sensor and will dominate every reading. Take the case off before you start.

Camera, microphone and Wi-Fi: not useful for vehicle trackers

Some detector apps advertise camera-based or network-based scanning. Those techniques belong to hidden camera detection, not tracker detection, and we cover their real limits in the hidden camera detector app guide. A GPS tracker has no lens to reflect light and, in the overwhelming majority of cases, no Wi-Fi radio to appear on a network. We found no research supporting any microphone or ultrasound based method of detecting a tracker, and would treat any app claiming one as unsupported.

What no app on any phone can do

This section is the reason this page exists. These are not limitations of current apps that a future update might fix. They are consequences of what a phone physically is.

It cannot sweep radio frequencies

A hardware bug detector is a wideband power meter: a broad antenna feeding a detector that measures total radio energy across a huge span of frequencies, with no ability to tell one signal from another. A phone is built to do the exact opposite. Each of its radios, cellular, Wi-Fi, Bluetooth, GPS, NFC and ultra-wideband, sits behind its own antenna and its own filter that deliberately throws away everything outside its assigned band. The hardware physically cannot deliver out-of-band energy to the receiver, and the basebands are closed systems that demodulate only their own protocol. There is no API on iOS or Android that returns raw wideband signal strength.

It cannot find a passive logger

A passive logger writes positions to internal flash memory and is read out later over USB. A GPS receiver only receives; it broadcasts nothing. The device therefore emits no radio energy at all and is invisible to every radio-based method, on a phone or on a thousand-dollar instrument. The NYU researchers who published the best recent work on tracker detection say so directly: their method does not work on passive trackers that store data locally. The only ways to find one are a physical search, and a magnetometer sweep at contact range if the housing happens to be magnetic.

It cannot catch a cellular tracker between bursts

This is the case most people are actually worried about, and there is now good published evidence on it. At USENIX VehicleSec in 2025, a team from NYU Tandon tested five consumer LTE trackers, including the Tracki, the Brickhouse Spark Nano 7, the Amcrest and the LoneStar Oyster3. What they found:

• Every device transmitted at fixed intervals, one minute for four of them and five minutes for the fifth. Between those bursts the devices are radio-silent and completely undetectable.
• Detection was done on the LTE uplink, at 709 MHz, 1877.5 MHz and 701.5 MHz depending on the carrier and band.
• The equipment was a tinySA Ultra spectrum analyser costing about $150, plus cell tower databases. Not a phone.
• Reliable detection range was three feet or less, with signal strength falling off sharply beyond that.

Read that list again with a phone in mind. Even with real test equipment and a published method, you have to be within a metre of the device and you have to be there during the one second per minute that it is talking. A phone cannot perform any step of that process.

It cannot see inside metal

Metal enclosures shield radio and redirect magnetic flux. A tracker inside a steel bumper, behind a wheel arch liner, or in the metal box section of a chassis rail is harder for every method, including the hardware ones. Even the manuals shipped with cheap RF detectors admit they cannot localise through structure.

The detection already built into your phone

Before you download anything, know that both major platforms already ship unwanted-tracker detection, and for Bluetooth tags it is better than what an app can do.

Apple and Google published a joint industry specification, Detecting Unwanted Location Trackers, on 13 May 2024. It is implemented in iOS 17.5 and later and Android 6.0 and later, and on both platforms the alert reads “Found Moving With You”. Chipolo, eufy, Jio, Motorola and Pebblebee committed to it at launch. A compliant tracker has to be detectable when separated from its owner, has to let you see its identifier, has to play a sound on demand, and has to give you disabling instructions.

On iPhone

• Alerts appear as “AirTag Found Moving With You”, “[Product Name] Found Moving With You” or “Unknown Accessory Detected”.
• You get Play Sound, and on ultra-wideband iPhones Precision Finding that walks you to the tag, plus instructions to disable it.
• Four things must all be switched on or you get nothing at all: Location Services, Find My iPhone, Bluetooth, and Tracking Notifications. Airplane mode must be off. It is worth checking these right now.
• Separately, an AirTag separated from its owner will start emitting a sound by itself, at a random interval between 8 and 24 hours since a change Apple made in June 2021.

On Android

• Works on Android 6.0 and later, and it covers Apple AirTags as well as Find Hub tags.
• Automatic background alerts, plus a manual sweep: Settings, then Safety & emergency, then Unknown tracker alerts, then Scan now. A manual scan takes about ten seconds.
• Google states the limits plainly: one alert per tracker per day, and alerts are deleted after 48 hours.

What the built-in alerts do not catch

The industry specification is a genuine improvement and it is also much narrower than the coverage most people assume it has. These are the gaps, and they are large.

1. Every cellular GPS tracker. The specification is a Bluetooth tag protocol. Tracki, Spark Nano, Oyster and the entire category of SIM-based trackers most commonly used on vehicles sit completely outside it. No alert will ever fire for one.
2. Non-participating tags. Unbranded Bluetooth tags from marketplace sellers are under no obligation to implement the spec.
3. Tampered tags. AirTags with the speaker removed were openly sold on eBay and Etsy. Apple added extra adhesive to the speaker module on AirTag 2 in February 2026 to make removal harder, which is a mitigation rather than a fix. Firmware modification is listed as a distinct threat in the IETF working group’s own threat model.
4. The person you live with. This is the most common real scenario and the specification handles it worst. The separation timer that triggers an alert resets every time the tag sees its owner’s phone. Geoffrey Fowler tested this for the Washington Post and found that when an abusive partner shares a home, the sound may never trigger at all. He also found the sound played for only about fifteen seconds and was easily muffled by putting an object on top of the tag.
5. A spare phone. An old handset in a glovebox with location sharing enabled is a tracker, and no tracker alert will ever mention it.
6. Your own accounts. Location shared from your own Apple, Google or Life360 account is not a tracker at all, it is a setting, and finding it means auditing your accounts rather than searching your car.

The twenty-minute physical search

This is the part that actually works, and it is free. A tracker has to be somewhere a person could reach in a hurry, and it has to either see the sky or sit behind non-metal bodywork. That narrows the search enormously. Work through the zones in order and do not skip the boring ones.

Before you start

• Take your phone out of its case if you plan to use a magnetometer app, and turn off any magnetic mount.
• Get a torch and something to lie on. Daylight is better than a garage.
• If you have any reason to think this is about a person rather than a company vehicle, read the safety section before you touch anything you find.

Zone 1: the OBD-II port, two minutes

Reach under the dashboard on the driver’s side, near the steering column. The diagnostic port is a trapezoidal socket. If something is plugged into it that you did not put there, that is a tracker, and it is the single most common placement because it needs no battery and takes ten seconds to fit. Look also for a small device spliced into wiring near the port.

Zone 2: the wheel arches and underside, seven minutes

This is where magnetic cases go. Run your hand around the inside lip of each wheel arch, then along the frame rails, the underside of the bumpers front and rear, and the tow hitch area if you have one. You are feeling for a box roughly the size of a deck of cards, often in a black or grey weatherproof case. This is the one zone where a magnetometer app earns its place: sweep slowly with the phone almost touching the metal.

Zone 3: the bumpers and body cavities, four minutes

Trackers are frequently placed behind plastic bumper covers rather than on metal, because plastic does not block the GPS signal. Feel up inside both bumpers as far as you can reach. Check the licence plate mounting, both sides.

Zone 4: the interior, five minutes

Under both front seats and in the seat rails. Inside the glovebox and centre console, including underneath the liner trays. In the door pockets. Under the floor mats. In the boot: lift the floor panel and check the spare wheel well, which is a favourite because nobody looks there for months at a time.

Zone 5: the obvious hiding places people forget, two minutes

The roof, if you have roof rails. Inside the fuel filler flap. The engine bay, along the inner wings. And any bag, jacket or child seat that lives permanently in the car, because a tag in a rucksack tracks you just as well as one on the chassis.

Hardware detectors: what the evidence actually says

The handheld detectors sold under names like K18, G318, G319 and Pro 10G cost roughly $20 to $200. It is fair to ask whether they are worth buying, and unusually for this field there is a real answer.

What is in the box

A typical G318-class unit claims a frequency range of 1 MHz to 8000 MHz, sensitivity better than 0.03 mW and dynamic range above 73 dB, from the manufacturer’s own datasheet. Note that its detection ranges are quoted in square metres of coverage rather than distance, which is a marketing unit rather than a physical one. Inside, there are usually three separate functions: a wideband radio power detector, a ring of red LEDs with a viewing filter for spotting camera lenses, and sometimes a magnetic probe.

Two of those three are exactly what a phone app does. The broadband radio stage is the only thing in the box that a phone genuinely cannot replicate.

The numbers nobody quotes

The most useful independent measurement in this field comes from a study presented at ACM SenSys in 2021, in which researchers built a detection method using a phone’s time-of-flight sensor and benchmarked it against alternatives with 379 participants. They tested the K18, one of the most widely sold detectors on Amazon, head to head.

In the only quantified independent comparison we could find, a purpose-built hardware detector missed roughly four hidden cameras in ten and raised a false alarm in a quarter to a third of trials. It still beat searching by eye. The authors also noted that detectors of this class can be difficult to use correctly.

Two caveats, stated because they matter. That benchmark measured camera detection, not GPS tracker detection, and we could find no equivalent peer-reviewed test of these devices against cellular trackers or audio bugs. And a detector’s own manual will tell you it alarms on your Wi-Fi router, your neighbour’s router, a phone call in the next room and a nearby cell mast, because it has no frequency selectivity whatsoever. The prescribed workaround is to move the vehicle to a weak-signal area, which for a parked car is often impossible.

So should you buy one

If you are checking a vehicle for a cellular tracker, a cheap detector will find it only during a transmission burst and only at close range, and it will alarm on a dozen innocent things in between. Your money buys more certainty in the form of twenty minutes of careful physical searching, or, if the stakes are high, a professional counter-surveillance sweep. If you want the genuine article rather than the consumer version, the NYU work used a tinySA Ultra spectrum analyser at around $150, which is a real instrument rather than an LED bargraph.

You found something. Now stop and read this

If the situation is not about a person, for example you have bought a used car and found a tracker left by a finance company or a previous fleet operator, none of the above applies and you can simply remove it. The distinction matters, which is why we are not telling everyone to do the same thing.

What we are building, and what it will not claim

Frequently asked questions

Is there an app that can detect a GPS tracker?

Partly. An app can detect Bluetooth tracking tags such as an AirTag, and it can help you find the magnets in a magnetic-mount tracker if you sweep the phone at contact range. No app can detect a cellular GPS tracker while it is silent between transmission bursts, and no app can detect a passive logger that never transmits at all. Those are limits of the phone’s radio hardware, not of the software, so no future update will change them.

Can my phone detect a hidden GPS tracker in my car?

It can detect one specific kind. Bluetooth tags are detectable and both iOS and Android now alert you to them automatically. Cellular trackers, which are the type most often used on vehicles, are not detectable by a phone. For those, a twenty-minute physical search of the OBD-II port, wheel arches, bumpers, seats and spare wheel well is far more effective than any app.

Do GPS bug detector apps actually work?

Some of what they do is real and some of it is not. Bluetooth scanning and magnetometer sweeping are genuine techniques with genuine limits. Any app claiming to sweep radio frequencies from 1 MHz to 8 GHz is claiming something the phone’s hardware physically cannot do, because each radio in a phone sits behind a filter that discards everything outside its own band and no operating system exposes raw wideband signal strength.

What is the best free way to check for a tracker?

Two things, in this order, both free. On Android, open Settings, then Safety and emergency, then Unknown tracker alerts, and run a manual scan, which takes about ten seconds. On iPhone, confirm that Location Services, Find My iPhone, Bluetooth and Tracking Notifications are all switched on, because if any of them is off you receive no alerts at all. Then do the physical search of the car described on this page.

Can a phone detect an AirTag?

Yes, and this is the one case where a phone is the right tool. Since iOS 17.5 and on Android 6.0 and later, both platforms alert you to a tracker that is moving with you and is separated from its owner, using the industry specification Apple and Google published in May 2024. The alert reads “Found Moving With You”, and you can play a sound to locate the tag. The significant gap is that the separation timer resets whenever the tag is near its owner, so if the person who placed it lives with you, the alert may never fire.

Why does my detector app alarm constantly?

If it is a hardware detector, because it has no frequency selectivity at all. It measures total radio energy, so your own Wi-Fi router, a neighbour’s router, a phone call in the next room, a smart meter or a nearby cell mast all trigger it. The manuals for these devices say so and advise moving to a weak-signal area, which is rarely possible with a parked car. If it is a phone app, it is most likely reading the magnetometer and reacting to the car’s own steel.

How much does a real GPS tracker detector cost?

Consumer handheld detectors run from about $20 to $200. In the only independent benchmark we found, a widely sold $30 class detector achieved 62.3% and 57.7% detection rates across its two modes with false positive rates of 26.9% and 35.2%, in a 379-participant study, for camera detection. Genuine test equipment is a different category: the NYU research team that published a working cellular tracker detection method in 2025 used a tinySA Ultra spectrum analyser costing around $150, and still needed to be within three feet of the device during a transmission burst.

What should I do if I find a tracker on my car?

If there is any chance this involves a person rather than a finance company or fleet operator, do not remove it yet. Photograph it in place, note the date and time, and speak to a domestic abuse or stalking advocate before you change anything. Removing it signals to whoever placed it that you know, and it destroys the evidence. Google’s own documentation warns that disabling a tracker may prevent law enforcement from identifying its owner. Full guidance and helpline numbers are in the safety section above.

Can a tracker work if my car is in a garage?

A GPS receiver needs a view of the sky to get a fix, so a tracker in an underground car park usually cannot determine its position. A cellular tracker will still typically report its last known position or attempt to connect when signal returns. A Bluetooth tag does not use GPS at all and reports its position through nearby phones, so it will keep working anywhere people walk past.

Does a faraday bag or signal blocker help?

It stops a device transmitting while it is inside the bag, which is occasionally useful for a phone or a tag you have already found. It does nothing to help you locate a device you have not found, and it does not affect a passive logger, which was never transmitting in the first place.

Is it legal to put a tracker on someone’s car?

That depends entirely on where you are and on the relationship between the parties, and it is a question for a lawyer rather than for us. What we can say is that in many jurisdictions placing a tracker on a vehicle you do not own, without the owner’s consent, is a criminal offence, and that finding one is grounds to contact the police. We are not lawyers and this is not legal advice.

Related reading

• Hidden Camera Detector App: What Works, What Doesn’t, and How to Search a Room · the same honest treatment for cameras, including the lens reflection and infrared techniques and what the research measured.
• Bug Detector App: Can a Phone Detect Listening Devices? · why audio bugs are the hardest of the three to find, and what actually works.

Sources

Satt et al., You Can Drive But You Cannot Hide: Detection of Hidden Cellular GPS Vehicle Trackers, USENIX VehicleSec 2025, NYU Tandon. Sami, Tan, Sun and Han, LAPD: Hidden Spy Camera Detection using Smartphone Time-of-Flight Sensors, ACM SenSys 2021. Heinrich, Stute, Kornhuber and Hollick, Who Can Find My Devices?, Proceedings on Privacy Enhancing Technologies 2021(3). Heinrich, Bittner and Hollick, AirGuard, ACM WiSec 2022. Apple and Google, Detecting Unwanted Location Trackers industry specification, announced 13 May 2024, and the IETF DULT working group documents including draft-ietf-dult-threat-model-05, 6 August 2026. Apple developer technote TN3111, iOS Wi-Fi API overview. Apple Support documentation on unwanted tracking alerts and Personal Safety. Google Android support documentation on unknown tracker alerts. Coalition Against Stalkerware survivor guidance and NNEDV Safety Net location tracking guidance. Manufacturer datasheets for the G318 and G319 detectors. Geoffrey Fowler, Washington Post, AirTag stalking test, May 2021. All accessed August 2026.

Apple, AirTag, Find My, Google, Android, Tile, Chipolo, Tracki and other product names are trademarks of their respective owners. Appsera is not affiliated with, endorsed by or connected to any of them. Product details, prices and platform behaviour described here were accurate in August 2026 and may have changed. Nothing on this page is legal advice.