Upload any MP3 or WAV file — or record live. Our model analyses 88 acoustic features to classify the voice in seconds.