Sonic: the fastest, most realistic text to speech model

Cartesia is the only model where you don't have to pick between quality and speed. Our models are built on State Space Model (SSM) architecture — a fundamentally different approach to transformers, pioneered by our founding team at Stanford. For text-to-speech, this translates into three things that matter at production scale: lowest latency in the market (Sonic delivers sub-90ms model latency, with streaming support that lets playback begin before the full response is generated); lower cost and higher concurrency (SSMs are computationally more efficient than transformers on long sequences); and state-of-the-art naturalness (higher accuracy on alphanumerics and heteronyms, and #1 ranked on third-party blind benchmarks for naturalness). Teams typically choose Cartesia when they've hit the limits of other providers on latency, reliability-at-scale, or deployment flexibility.

Yes, and this is one of the main reasons enterprise and government buyers choose Cartesia. Sonic-3.6 can be deployed on-prem inside your data center (including air-gapped environments), in your own VPC on AWS/GCP/Azure, or via OEM licensing for embedding Sonic directly into your product. This makes Cartesia viable for government contracting, regulated industries (healthcare, financial services, insurance), and customers with data sovereignty or residency requirements. On-prem and OEM deployments are available under enterprise contracts.

Cartesia is built for enterprise and regulated industry deployments. Our compliance posture includes SOC 2 Type II, HIPAA-eligible (with BAAs available for healthcare customers), GDPR-compliant, zero data retention available for enterprise customers across eligible services, and on-prem/VPC deployment for customers with strict data residency, sovereignty, or air-gapped requirements. Our Data Protection Addendum can be found at https://www.cartesia.ai/legal/dpa.

You can clone voices you have the right to clone. Cartesia offers Instant Voice Cloning from a short reference sample (under a minute of clean audio), Professional Voice Cloning from longer reference audio (15–30 minutes) for the highest-fidelity clones in production, and custom voice development under enterprise contracts for customers building branded voices at scale. All voice clones require verified consent from the speaker. Cloning voices you don't have permission to use — including public figures, celebrities, or other people without their consent — is prohibited under Cartesia's Terms of Use. Cloned voices work across Sonic TTS, the API, and Managed Agents.

Cartesia offers plans across three tiers. Free: for evaluation and personal use, includes 20K credits and access to Sonic, the API, voices, and Instant Voice Cloning. Paid self-serve plans: for teams shipping to production, includes commercial usage rights, higher credit allotments, higher concurrency, and faster generation; pricing is at https://www.cartesia.ai/pricing. Enterprise: for high-volume production, on-prem and VPC deployment, OEM licensing, custom voices, zero data retention, and higher SLAs, with custom pricing based on usage and deployment model. Cartesia also offers a startup grants program at https://www.cartesia.ai/startups with credits for qualifying early-stage companies.

Reach out to the Cartesia team if you're running high-volume production workloads (>50M credits); you need on-prem, VPC, or OEM deployment; you need a BAA, zero data retention, or other contractual compliance terms for healthcare, financial services, or regulated sectors; or you're in government, federal, or public sector procurement. For everything else — evaluation, prototyping, smaller production workloads — the self-serve plans on the pricing page will get you started.