birdnet-go

Why we included this project

A Raspberry Pi paired with a cheap microphone becomes a 24/7 wildlife monitor: BirdNET-Go listens to a soundcard or a network audio stream, classifies what it hears with AI models running locally on the device, and shows detections in a web interface you can open from any browser. The part that stands out is the model gallery, which lets you install the BirdNET bird classifier, Google Perch, or regional bat models from inside the app without recompiling, and run several at once on different audio sources so the models can confirm each other's detections. That combination of always-on local inference and privacy appeals to conservation groups, citizen scientists, and hobbyists who want continuous monitoring without uploading audio to the cloud. Developers evaluating edge AI on constrained hardware also get a compact Go implementation worth studying.