AI Hub

Sound Recognition

Identify environmental sounds and audio events in real time using AI — runs entirely in your browser, no upload needed.

Press Start to begin sound recognition.

All audio is processed locally in your browser. Nothing is uploaded.

How to use

  1. Click Start and allow microphone access.
  2. Hold the microphone near the sound source.
  3. The AI labels the top sounds in real time.

Examples

  • Identify music genres or instrument types.
  • Detect environmental sounds like rain, traffic, or birds.
  • Recognise speech, laughter, or other human sounds.

About the Model

Source: google/yamnet

License: Apache 2.0

521-class AudioSet classifier — loaded via TFHub (tf.loadGraphModel + fromTFHub). All audio processed on-device.