Hume AI

Hume AI trains speech and multimodal models designed to understand and generate human emotional expression, powering low-latency conversational voice systems. It released its Empathic Voice Interface in 2024 and shipped successive expression models through 2026.

Ecosystem functionExplicitly models prosody and emotion rather than optimizing transcription accuracy or naturalness alone.

Business modelVoice model APIs