appAves — a bioacoustic station inside a phone

Monitoring birds in the field usually calls for expensive equipment. appAves takes another route: it turns a reused Android smartphone into the computer of a field station. It captures audio through a standard USB converter, identifies the species right there —sending nothing to the cloud— and stores the event together with its sound clip.

It is built to spend months out in the wild on solar power: it runs as a foreground service, with its own watchdog and health telemetry so you can tell remotely whether the station is still alive.

Technically it is native Kotlin with C++ layers over the NDK, organised in modules with a hardware abstraction layer so it can be ported between different phones. Species identification runs on TensorFlow Lite.

An honest note about the identification model: the recognition model it uses is distributed under a non-commercial licence and is not bundled with the project — the app downloads it on first launch. Commercial use requires agreeing terms with its authors, or replacing it.