Hume AI

Hume AI is a research-driven voice-AI company whose distinctive edge is emotion. Its flagship product, the Empathic Voice Interface (EVI), is a low-latency voice-to-voice model that listens to your prosody, detects the emotional state behind your words, and responds with matched expressive speech — you can build an agent whose tone actually changes with the conversation. EVI 3, released in 2026, added a fully customizable voice studio, cross-language expressiveness, and tool-use so the voice agent can call your APIs or MCP-style tools mid-conversation. Hume also ships Octave, a high-quality expressive TTS engine with fine-grained emotion and pacing controls, plus a separate Expression Measurement API for classifying vocal, facial, and text emotion from any input.

The stack is delivered as REST and WebSocket APIs, with SDKs for Python, TypeScript, iOS, Android, and React Native. Pricing is usage-based with a small free tier for prototyping, then per-minute or per-character billing on voice, and per-request billing on Expression Measurement. Enterprise contracts add custom voice creation, on-prem or VPC deployment, and BAA/HIPAA. Hume is the go-to voice provider for teams building emotionally attentive tutors, mental-wellness apps, cars, robots, and consumer companion products — anywhere the how of a spoken answer matters as much as the what.