How detection works

Every detection method QuietCheck uses is heuristic and approximate. None of them are a guarantee. This page explains, in plain language, how each one actually works and where it falls short.

The short version QuietCheck tells you something is probably nearby, with a stated confidence level. It never tells you something is definitely there, and it can miss real devices or flag things that turn out to be harmless.

Smart glasses, trackers, and drones — Bluetooth signature matching

Bluetooth Low Energy (BLE) devices continuously broadcast small "advertising" packets that anyone nearby can listen to. That's standard, publicly visible Bluetooth behavior; QuietCheck isn't breaking into or bypassing anything. QuietCheck checks each broadcast's manufacturer identifier and, where relevant, its advertised name, against a list of identifiers known to belong to specific smart-glasses, tracker, or drone hardware makers.

For drones, where manufacturer-ID matching doesn't apply, QuietCheck instead parses the standardized ASTM F3411 Remote ID broadcast that most drones now transmit over BLE, a real aviation-safety standard.

Every match gets a confidence tier instead of a flat yes/no:

High — a manufacturer ID verified against an authoritative registry, from a company whose only real product line matches the category (e.g. a dedicated Bluetooth-tracker maker), or a validly-formatted Remote ID broadcast.
Medium — a real, verified manufacturer ID, but from a company whose product line is broader than just this device category, so a match is a weaker signal.
Low / Unverified — a plausible but unconfirmed signal, shown so you're not left with no information at all, but clearly labeled as needing more confirmation.

What this can't do:

Identify a specific individual device with certainty. A device can share a manufacturer identifier with a piece of matching hardware without actually being one.
Determine a real compass bearing or exact distance. Distance estimates come from a standard log-distance path-loss model based on signal strength, which is a noisy measurement to begin with: multipath reflection, device orientation, obstructions, and power differences between devices all throw it off. The angle shown on QuietCheck's radar view is illustrative, not a real bearing; BLE doesn't support direction-finding.
Detect a device that isn't broadcasting BLE at all, or that broadcasts under an identifier QuietCheck doesn't yet recognize.

Hidden cameras — camera-and-flashlight "lens glint" detection

QuietCheck's optional Hidden Camera Check uses your device's camera and flashlight to look for a specific visual pattern: with a flashlight on in a dim room, a camera lens reflects light back as a small, sharp, unusually bright point, distinct from the softer, larger reflections of ordinary surfaces. QuietCheck looks for that pattern in real time, on-device.

It's a heuristic brightness-and-contour check rather than a trained machine-learning model, the same general technique other hidden-camera-detector tools use.

This is a heuristic aid, not a guarantee. It will miss real hidden cameras plenty of the time: wrong angle, no reflection at the moment you scan, ambient light too bright, a lens too small or recessed. It can also flag things that aren't cameras at all, like jewelry or glass. Don't rely on it as your only way to check a space for hidden cameras.

Video is analyzed frame by frame, in memory, in real time. Nothing from this feature is ever recorded, saved, or transmitted anywhere. See the Privacy page for the full detail.

What QuietCheck will never claim

QuietCheck doesn't claim to detect every smart-glasses model, every tracker, every drone, or every hidden camera, and it doesn't claim perfect accuracy for anything it does detect. Background scanning (continuing to listen after you leave the app) is inherently best-effort on iOS, so a lack of an alert while backgrounded doesn't mean nothing is nearby.