The BirdNET team’s setup for the MacOS command-line Python version is on GitHub here.

Here is a short step-by-step that I wrote for myself (a MacOS user – YMMV)

These steps are now very close to those found on the BirdNET-Analyzer GitHub page, but they have a few more details and reminders-to-myself.

Use Python 3.11 (click here for Python-install reminders)

CD into a directory you’ve created to house BirdNET-Analyzer.  Pay attention to what you name it because that will become part of the path that’s used to call the script.  For example; avoid spaces in the directory name.  Here’s an example you can use:

cd BirdNET-Analyzer_GitHub_clone

git clone https://github.com/BirdNET-team/BirdNET-Analyzer.git (or the latest "clone" URL from GitHub)

CD down into the newly-cloned BirdNET directory.  It should be the only thing the directory you just created.

cd BirdNET-Analyzer

From inside the BirdNET-Analyzer directory, create and activate a new virtual environment (it’s named “venv” in this example).

python3 -m venv venv

source venv/bin/activate

Install BirdNET’s requirements manually (yes, really use a period at the end…) :

pip install .

Test by running the help command:

python3 -m birdnet_analyzer.analyze --h

Verify the installation

python -m birdnet_analyzer.analyze <path_to_an_audio_file_on_YOUR_drive.wav>

Here’s an example of a command-line that worked on my Mac the day I wrote this.  Note that there’s no longer a parameter for the input-directory path (ala “-i”).  It’s positional; it’s always the last entry in the command.

python3 -m birdnet_analyzer.analyze --threads 8 --lat 44.415947 --lon -91.700659 --skip_existing_results -o /Users/mikeoconnor/Documents/birdnet_analyzer/run_files/2025 /Volumes/Bio_2025

Translated, that command is saying:

python3

— run this Python script

-m

run a module of a file rather than a file

birdnet-analyzer.analize

the name of the script and the module to run

--threads 8

let the script use up to 8 parallel processes on my machine (YMMV)

--lat and --lon

geo coordinates of the recordings (narrows down the analysis)

--skip_existing_results

only process new files in the directory

-o

path to the destination directory

/Volumes/Bio_2025

the last chunk is the path to the input directory, it doesn’t have a parameter it’s positional