whoBIRD

Why we included this project

whoBIRD takes BirdNET's acoustic recognition models and turns them into an Android app that works offline, so identifying a song from a bush doesn't require a signal. It listens through the phone's microphone, matches calls against a database of more than 6,000 species, and keeps a running log of everything you've heard, which you can back up or share. The on-device processing means it keeps working in remote spots, and the settings are tuned for actual fieldwork: a high-pass filter for traffic noise, an adjustable detection threshold, and a location- and date-aware meta-model that helps pin down migratory species around the seasons. It's a good example of a research model made practical for anyone who carries a phone, whether you're a serious ornithologist or just curious what's singing outside the window.