Hume
Expressive AI voices and a voice assistant that reads tone
- Made by
- Hume AI
- Category
- Voice
- Last change
- Nov 2, 2026
Changelog
November 2026
- Launch
Hume launched the Audio Decision API, which answers questions about speech and sound, such as speaker count, laughter or caller frustration, in a single call for developers.
What this means for you: Hume's Audio Decision API lets developers send a recording and ask a question about it, such as whether the customer sounds frustrated, how many people are speaking, or where someone laughs. It returns an answer with a confidence score. One example is searching an archive of 40,000 recordings for every call where the customer sounded frustrated.
How to use it: Request API access by joining the waitlist on Hume's site. You can try live and uploaded-file audio in the playground, where you add your own questions. The docs list supported questions.
Related tools
- SiriApple's assistant on iPhone, iPad and Mac, rebuilt with AI as Siri AI (in beta).
- SuperwhisperVoice-to-text dictation for Mac and iPhone that types into any app
- ElevenLabsAI voices, dubbing, music and sound effects
- KrispNoise cancellation, meeting notes and accent conversion for calls
- Alexa+Amazon's new generative-AI Alexa that can take actions for you
- Fish AudioAI voices and voice cloning from a short audio sample.
