Standout features
Hume AI builds emotionally intelligent voice — its Octave TTS interprets meaning to deliver expressive speech, and EVI listens for emotion and responds empathetically in real time.
Worldwide search interest, indexed 0–100 · Google Trends.
Hume AI focuses on emotional intelligence in voice — Octave TTS reads the meaning behind text for expressive delivery, and EVI detects the user’s emotion and responds empathetically.
- New York lab; founder Alan Cowen (ex-DeepMind, emotion science).
- Octave TTS: LLM-based, interprets context for expression.
- EVI: real-time empathic voice that reads + mirrors emotion.
- Works as a voice layer for any LLM (Claude, GPT, Gemini).
Hume is built around emotion.
- Octave TTS — contextual, expressive speech.
- EVI — speech-to-speech that detects + responds to emotion.
- Expression-measurement models (vocal/facial/verbal).
- Voice cloning + design-by-prompt; multilingual.
Freemium, usage-based.
Hume fits empathy-driven apps.
- Companions + wellness / coaching apps.
- Empathetic customer-support agents.
- Interactive experiences where tone matters.
- Pure low-cost batch narration.
- Teams wanting the biggest stock voice library.
No tool is perfect — the trade-offs to weigh:
- Niche focus — emotion over raw breadth.
- Developer-oriented APIs, not a creator suite.
- Smaller than ElevenLabs / Google.
- Usage pricing to monitor at scale.
- ✓Reads + responds to emotion
- ✓Context-aware Octave TTS
- ✓Works with any LLM
- ✓Voice cloning + design
- ✓Strong for companions / support
- ✕Niche emotion focus
- ✕Developer-oriented APIs
- ✕Smaller than leaders
- ✕Usage pricing to monitor
Developers value Hume for emotionally aware voice — Octave’s context-driven expressiveness and EVI reading and mirroring the user’s emotion, usable with any LLM. The gripes are its niche emotion focus, developer-oriented APIs, and smaller scale than the leaders. Sentiment is positive for empathy-driven apps.
Hume AI is a New York lab building emotionally intelligent voice AI.
Company figures are drawn from public disclosures and reputable trackers (gathered Jun 2026). User and revenue numbers are estimates and move fast.
Pick up to two other coding tools to see them head-to-head on the same rubric.