Hume launches an API that answers questions about speech and sound
Hume AI introduces its Audio Decision API, an audio-native model that returns an answer with a confidence score for questions the user defines.
By GennnAi··1 min read
The model analyzes raw audio without transcription, for example caller frustration, AI-generated voices or a siren in the background. It can query live calls and make an archive of 40,000 recordings searchable in minutes. In an initial evaluation, it reached 70.7% agreement with a human panel's average judgment, versus 56.5% for an individual rater. Hume claims better accuracy, speed and cost than existing methods. Access requires joining a waitlist.
Source: Hume AI, published on November 2, 2026.