Hume Launches Audio Decision API Answering Voice Questions in One Call
EXM7777 · x · 2026-11-03
Hume AI introduced the Audio Decision API, letting you ask natural-language questions about audio: how many speakers are present, whether anyone is laughing, if a caller sounds frustrated, and more.
Trained on over 10 million human listening judgments, it answers the way human listeners would. The company says tasks that previously required chaining 10+ models now take a single API call, with support for both realtime streams and hours-long recordings. Access is currently via request.
Related event: Hume Launches Audio Decision API for Native Audio Understanding(3 posts)→
More from Multimodal
- Claymation-style blue-and-pink AI-generated botanical portraits go viral — gizakdag · 2026-11-03
- Grok explains the etymology of 'kumi-daiko' while composing taiko music — TinfoilTricorn · 2026-11-03
- GANPaint, the classic mechanistic interpretability tool, revived via vibe coding — hen_str · 2026-11-03
- AI-generated 'Nike Shades' concept with swoosh devil horns is absurdly convincing — Delahuntagram · 2026-11-03
- AI artwork 'Digital World' lands Midjourney Magazine Issue 42 poster — ciguleva · 2026-11-03
- Runway showcases AI short film "Catch Me," made by one creator in days — runwayml · 2026-11-03