Shadowband
Android · Kotlin · built, shipped and maintained solo
Over 7,000 installs · 4.2 average rating — September 2026
Why I built it
The paranormal has always been something that’s interested me, and I’ve when I’ve tried apps and ITC devices in the past, I’m never fully sure exactly what’s happening under the hood. Why is it saying X?, What is happening to trigger the sound?, Most importantly, Am I being tricked here? I wanted to know what was happening underneath. In building Shadowband, I understand this a little more, but, honestly, I’m still not sure exactly what’s happening beyond the app. The app uses environmental readings, such as temperature, EMF, sound, proximity, vibration, even camera in some modes. I know now what the app is doing, but I’m honestly still not sure some times what the environment is :D I get that electronics can impact EMF and cause fault positives, which I tried to rule out in the app, but as fo rother fluxuations, what exactly causes them? the weather? air moving around a building? Earth tremors? Ghost? You’re guess is as good as mine :)
I wanted to create an app that was free and fun for people to use, not something full of adds, or where you’re having to pay for each module. So, I created Shadowband :)
What it does
The app has various modes:
The classic ITC (Instrumental Trans Communication) or Spirit Box - Shadowband.
Camera enabled visual figure mapping, using AI assist (Mediapipe/ML_Kit) - Apparition Scanner.
A REM Pod and trigger object/image movement sensor combined - Trigger REM Pod.
I plotting grid for highlighting areas of interest - Energy Map
Somewhere to view all your logs and captures - Evidence Locker
The interesting technical problem
The latest addition to the app is low frequency audio detection, which is more difficult than it sounds on consumer hardware. Phone microphones aren’t designed for it, so getting anything trustworthy out of one means:
- Heterodyning to shift infrasound up into the audible range
- 50Hz notch filtering to remove mains hum, which otherwise dominates everything at the low end inside any building
- YAMNet TFLite classification, gated behind an adaptive spectral baseline so it responds to genuine changes rather than firing constantly on ambient noise
Other recent work: a camera crash fix, a BroadcastReceiver leak fix, a Fahrenheit/Celsius toggle, and a High Gain mode.
Interest and coverage
- Newton News — Covered the app in the October 2025 edition.
- Creepy Blinders Used by a local paranormal Group on investigations.
