BirdNET-Pi acoustic bird monitor

What you’re building

BirdNET-Pi continuously records audio from a USB microphone or sound card, runs BirdNET-based machine-learning classification, extracts likely bird clips, and publishes detections, trends, a live stream, spectrogram, logs, and administration tools through a local web interface. The original source supports Raspberry Pi 4B, 400, 3B+, and Zero 2 W with 64-bit Raspberry Pi OS; the lower-powered 3B+ and Zero 2 W require the Lite image. The expected outcome is a 24/7 acoustic bird-monitoring station reachable at birdnetpi.local, with optional BirdWeather sharing and automatic disk-space management. Two constraints materially change the editorial framing: the mcguirepr89 repository now declares itself deprecated and directs new installations to Nachtzuster’s maintained fork, and its license prohibits use to develop a commercial product. It also does not support installation on top of other server stacks, so a dedicated OS deployment is the documented path.

Key steps

1. Start from a supported Raspberry Pi and a fresh 64-bit Raspberry Pi OS image; use the Lite image for a Pi 3B+ or Zero 2 W.
2. Attach and confirm the USB microphone or sound card before installation.
3. Follow the original README’s direction to the actively maintained Nachtzuster fork and run that fork’s newinstaller.sh on the dedicated system.
4. After installation, open birdnetpi.local or the Pi’s address, immediately configure the initially empty web password in Tools > Settings > Advanced Settings, and verify live audio.
5. Confirm detections, clips, spectrogram, and disk management, then configure species/location settings and optional BirdWeather sharing while respecting the non-commercial license.