Consumer privacy app combining magnetic, optical and network detection for hidden cameras
A native iOS utility that helps people check a space for concealed recording devices using three independent detection methods — magnetic sensing, camera-based optical analysis and local network scanning — rather than the single unreliable method most competitors rely on.
Most apps in this category implement one detection method, usually a magnetometer presented as a spinning dial, which reacts to any metal object and offers no way to corroborate a result. That makes them entertaining rather than trustworthy, and a false positive that fires constantly teaches the user to ignore the app entirely — which means it fails exactly when it matters. The person using it is also, by definition, already anxious and not technical, so raw sensor output isn't a usable result on its own; without proper interpretation the tool does more harm than good.
Built three independent detection paths rather than one, on the premise that any single sensor reading in this domain is ambiguous — magnetic sensing, camera-based optical analysis of lens reflections, and local network host discovery for connected devices that are neither magnetically nor optically detectable. Because the failure modes of the three methods don't correlate, agreement between them carries real information in a way no single reading can.
The optical analysis path processes camera frames per frame, which isn't a workload to leave in a higher-level layer on older hardware, so I dropped into C++ for that piece specifically, bridged into the interface through Objective-C++. Around the detection core, interpretation was treated as a first-class feature rather than a help screen — onboarding, guidance content and a photo capture-and-review flow so a momentary reading becomes something the user can act on, keep or show someone. Monetisation combined advertising for the occasional user with subscription for the frequent one, since a utility used a few times a year doesn't convert well on subscription alone.
Results can be corroborated instead of resting on a single ambiguous signal, and the app detects a class of device — network-connected cameras — that sensor-only competitors can't find at all. Because the demanding analysis runs at the right level, the app stays usable on older hardware, and because interpretation was built as a product feature, users come away with a decision rather than a number they don't know how to read.
Something holding you back?
New build, rebuild, or adding AI to what you already have — I'll take a look and give you an honest perspective.