Provider pricing

Cartesia AI API pricing

Cartesia builds the Sonic family of state-space TTS models — among the lowest-latency voice models available, with sub-90ms time-to-first-byte. The API covers voice cloning, multilingual TTS across 15+ languages, and a streaming endpoint purpose-built for voice agents and real-time apps.

Founded2023
HeadquartersSan Francisco, CA
Tracked modelsDirectory

Best for

Real-time voice agents, gaming NPCs, and any product where TTS latency defines UX.

Strengths

  • Sub-90ms TTFB on Sonic — lowest-latency TTS in production
  • Instant voice cloning from a few seconds of audio
  • State-space architecture: efficient, deterministic, and steerable
  • Multilingual support across 15+ languages