Common questions

Audio is processed in real time and optionally stored in your encrypted cloud account so you can review history and generate reports. You can delete all data at any time from Settings.

Two honest answers. Bark detection (is this a dog bark?) is measured — recalibrated on real field recordings. Emotion is harder: we measure it against a real labelled dataset, and from audio the energy/arousal signal is weak-but-real while the positive-vs-negative read is close to a guess. That's exactly why the mood axis comes from an optional camera read of body language (measured at AUC 0.93), not the sound. So treat an audio-only emotion as an informational acoustic estimate, not a clinical diagnosis — add a photo or video for the mood, and your feedback improves it over time.

Bark detection analyses acoustic features (pitch, rate, harmonics-to-noise) present across all breeds. The emotion model has so far been measured on two breeds (husky and shiba), so for other breeds emotion readings lean on breed-relative acoustic rules and should be treated as estimates. Detection works regardless of breed.

Silence detection tracks quiet periods too — useful for separation monitoring. You can also upload an existing clip instead of recording live.

It's free. Every feature, no subscription, no card required — just sign up and start recording.