Hume AI
Expressive voice APIs: real-time speech-to-speech (EVI) and instructable text-to-speech (Octave)
Hume AI offers two voice APIs. EVI (Empathic Voice Interface) is a real-time speech-to-speech model that reads vocal expression and adjusts its replies. Octave is a text-to-speech model steered with natural-language direction, with voice cloning and voice design. SDKs cover Python, TypeScript, React, Swift and .NET.
In January 2026 Google DeepMind hired founder Alan Cowen and several engineers under a licensing deal. Hume kept operating; its homepage now leads with voice AI evaluation (Kairos, Human Feedback API). The free tier includes 10,000 TTS characters a month as of September 2026.
Pricing: Free / monthly subscriptions
Hume AI Alternatives
Explore 24 products in the Audio category. View all Hume AI alternatives.
LiveKit Agents
Open-source framework for building real-time voice and multimodal AI agents over WebRTC
SLNG
Voice agent infrastructure that routes across STT, TTS and LLM providers with region-pinned execution
Fish Audio
Text-to-speech and voice cloning API, with Fish Speech weights published under a research licence
Speechmatics
Enterprise speech-to-text API supporting 55+ languages with high accuracy
Gladia
Fast speech-to-text API with real-time transcription and speaker diarization
Work on Hume AI? Feature it at the top of Audio.
Is your product missing?