Cartesia is a voice AI company that provides one API for speech generation, transcription and production voice agents fast enough for live conversation. Sonic is its text-to-speech model, Ink handles streaming speech-to-text, and Managed Agents lets teams build and ship enterprise voice agents. The models are built on state space models, a research direction the team helped pioneer, and the company reports first place on the Artificial Analysis speech leaderboards. A free plan includes 20K credits a month, and paid plans start at 5 dollars a month.
Developers building voice products and enterprises in finance, healthcare and government are the main audience. A startup can add realistic speech to an app with the API, a bank can run outbound verification calls on suspicious transactions, and a healthcare operator can deploy voice agents under strict data residency rules because inference can run in-region, on-premise or on-device.
Concrete uses include real-time assistants, call center automation, narration and voiceover, and live transcription. Instant voice cloning arrives on the Pro plan and professional cloning on Startup, while Scale adds priority support and higher concurrency, and Enterprise adds SSO, DPAs and BAAs for compliance. Voice agents are billed at 6 cents per minute, with telephony at 1.4 cents per minute on Cartesia-provided phone numbers. Every plan includes unlimited workspace seats, so teams of any size can share one account, and a cost calculator recommends a plan from expected minutes of generated audio, transcription hours and agent usage. Dedicated voice and language catalogs help teams choose a speaker for each market.
