Directory / Speech APIs (STT/TTS)
Cartesia
Ultra-low-latency, expressive voice models for real-time AI
Visit Cartesia ↗Who it's for
Voice agent developers, enterprises
Review: pros & cons
Cartesia builds real-time voice models known for very fast, expressive text-to-speech that suits conversational agents. Its pro is extremely low time-to-first-audio, which makes agents feel responsive. Limitation: it focuses on the speech layer, so teams still need separate orchestration, telephony and language models to build a complete voice agent.