Bioacoustic field recorder

Record animal sounds with evidence you can inspect.

AnimalEar turns phone recordings into live acoustic traces, candidate detections, and saved analysis views for birds, insects, pets, and field recordings.

AnimalEar app icon

Built for recording-first use on the web now, with native iOS and Android testing underway.

What the app does

Record first

Open the recorder quickly, capture audio, review a clip, and save evidence before adding more context.

Live acoustic clues

Waveform, waterfall, field, and live detection views show what the microphone is hearing while recording.

Cautious AI analysis

Candidate labels are presented as acoustic evidence, not guaranteed identifications or literal translations.

How the acoustic model works

One compact listener

The model turns each short window of sound into a small acoustic fingerprint, then compares that fingerprint against known sound patterns.

Taxonomy-aware results

Animal matches are tied back to AnimalEar taxonomy, so a recording can connect to species pages, images, notes, and future reviewed observations.

Objects and events

The same system can recognize non-animal sound clues such as alarms, engines, aircraft, water, glass, tools, and household alerts without mixing them into animal taxonomy.

Confidence and abstention

When the sound is distant, noisy, overlapping, clipped, or unfamiliar, AnimalEar can return a broader clue instead of forcing an exact species or object.

Expandable sound memory

Exact species, event, object, and private user-confirmed sounds can be added as small prototype packs without making the core phone model heavy.

Local first, cloud when useful

The phone should handle common real-time clues. Longer recordings, rare species, hydrophone audio, and messy soundscapes can be sent for deeper analysis when enabled.

AnimalEar does not publish private training sources or user recordings. Public explanations describe the model behavior, confidence, and evidence trail, not the exact training recipe.