Cartesia Sonic

Cartesia Sonic is a flagship text-to-speech (TTS) model that delivers ultra-low latency voice generation with a time-to-first-audio of just 90ms. Designed for fluid, real-time voice AI experiences, Sonic powers voice agents, customer service applications, localization, and interactive conversational systems. The platform includes Sonic-3 as the flagship TTS model, along with Ink-Whisper for speech-to-text and Line for voice agent development.

• Ultra-Low Latency TTS: Sonic-3 delivers industry-leading 90ms time-to-first-audio, enabling natural real-time conversations and voice interactions without perceptible delays.

• Voice Cloning: Offers both instant voice cloning (available on Pro and above) and professional voice cloning (Startup and above) for creating custom voice profiles with high fidelity.

• Voice Changer: Transform audio with voice modification capabilities, allowing users to alter voice characteristics in real-time.

• Multilingual Support: Comprehensive language support for global applications including localization across Asia Pacific, Europe, Latin America, Middle East, and more.

• Voice Library: Access a curated collection of pre-built voices for immediate use in applications without custom training.

• Design a Voice: Create custom voice profiles tailored to specific brand requirements and use cases.

• Infilling: Advanced text infilling capabilities for seamless audio generation with natural transitions.

• Line Voice Agent Platform: Build voice agents from first agent to production-ready deployment with SDK, CLI, telephony integration, call analytics, and observability tools.

• Ink Speech-to-Text: Ink-Whisper provides the fastest streaming speech-to-text at competitive pricing, complementing the TTS capabilities for full voice AI workflows.

• API Access: RESTful API with concurrent request support scaling from 2 (Free) to custom limits (Enterprise) for TTS operations.

To get started, sign up for a free account to receive 20K credits for models and $1 prepaid for agents. Upgrade to Pro for commercial use and instant voice cloning, or choose Startup/Scale plans for team collaboration, higher concurrency limits, and professional voice cloning capabilities.