AI-powered bioacoustics, at scale.

BirdNET uses deep learning to recognize over 6,000 species globally. Our open-source tools provide robust workflows for large-scale biodiversity monitoring and citizen science.

A joint effort by:

The AI toolkit for conservation.

Open-source models and interfaces designed to scale from individual observations to continental monitoring programs.

Biogeographical Priors
Refine species lists using precise coordinates and date metadata.
High-Throughput CLI
Optimized headless analysis for massive acoustic deployments.
Embedded Models
Lightweight model versions for low-power edge computing or mobile devices.
Open Source & Community
Contribute code to our core repositories or join our effort by sharing your data.

For Ecologists

Process months of PAM data with high precision. Filter results by location and time to reduce false positives.

For Developers

Integration-ready Python modules and TFLite models. Access our codebase and community-driven tools.

Citizen Scientists

Contribute to global biodiversity monitoring by using the BirdNET app to record and identify birds in your area.

Bioacoustic Research

Scientific publications detailing the models, validation studies, and ecological impact of BirdNET.

How BirdNET works

BirdNET processes raw acoustic data through a multi-stage pipeline designed for ecological accuracy. By transforming varying soundscapes into standardized feature representations, the system can isolate and identify subtle vocal signatures across thousands of species while accounting for biogeographical variations in species distribution.

1. Capture

Audio is captured at 48 kHz and divided into 3-second segments, optimizing the balance between model input size and the natural duration of avian vocalizations.

2. Spectrogram

Signals are processed into two log-scaled Mel-spectrograms, visualizing frequency patterns from 0 to 3 kHz and 150 Hz to 15 kHz for detailed analysis.

3. Neural Net

A Convolutional Neural Network (CNN) scans these visuals, utilizing millions of trained weights to detect species-specific patterns.

4. Result

Initial predictions are cross-referenced with local metadata (location and date) to produce high-confidence species probabilities.

A complete stack for research.

BirdNET is more than a model—it's an integrated ecosystem of tools designed to fit into your existing research workflow, from raw audio processing to statistical modeling in R.

birdnet (Python Package)

Native Python integration for developers and data scientists. Build custom pipelines or run large-scale inference on your own infrastructure.

birdnetR

The bridge to R-based ecological analysis. Clean, filter, and summarize BirdNET outputs directly within your R environment for occupancy modeling and trend analysis.

birdnetTools

User-friendly utilities for researchers. Access graphical interfaces for managing thousands of detections, verifying species, and preparing data for publication.

# Python Integration
import birdnet
model = birdnet.load("acoustic", "2.4", "tf")
predictions = model.predict("example/soundscape.wav")
predictions.to_csv("example/predictions.csv")

# R Ecological Analysis
library(birdnetR)
model <- birdnet_model_tflite()
audio_path <- system.file("extdata", "soundscape.mp3", package = "birdnetR")
predictions <- predict_species_from_audio_file(model, audio_path)
get_top_prediction(predictions)

Built with BirdNET

BirdNET powers a diverse ecosystem of external tools, from specialized field hardware to global data visualization platforms.

HaikuBox logo

HaikuBox

Acoustic Hardware
BirdNET-Pi logo

BirdNET-Pi

DIY Monitoring
BirdWeather logo

BirdWeather

Global Dashboard
Chirpity logo

Chirpity

Desktop Analysis
BirdNET-Go logo

BirdNET-Go

CLI & Integrations
ecoSound web logo

ecoSound

Web Monitoring

A peer-reviewed foundation for bioacoustics.

BirdNET is built on a foundation of rigorous research. Our work focuses on developing reliable, open-source tools that transform raw audio into peer-reviewed ecological data.

Fire Ecology

Our research in the Sierra Nevada mountain range evaluates how forest restoration and fire regimes affect bird occupancy, helping land managers understand biodiversity recovery in fire-adapted ecosystems.

Community Dynamics

We quantify how survey design, spatial coverage, and recording duration influence species richness estimates, providing a rigorous benchmark for monitoring shifts in avian community composition.

Research & Education Outreach

We provide tools and documentation for K-12 programs and workshops, supporting the next generation of researchers in bioacoustics.

Support the Future of Bioacoustics

BirdNET is an open-source research project powered by global collaboration. Whether you're a developer contributing code or a birder recording sounds, your help is vital.