Streaming speech transcription
Convert incoming audio to text with Ink-2 for live spoken applications.
Speech models and managed voice agents for teams building real-time voice experiences.
Cartesia provides speech-generation, transcription, and managed voice-agent services through a single API and product platform. Its Sonic text-to-speech and Ink speech-to-text models are designed for streaming and real-time interaction, while Managed Agents supports building and deploying voice agents. The platform is intended for teams adding spoken interfaces to applications and operational workflows. Cartesia describes deployment options across regional cloud endpoints, on-premise environments, and on-device use, with deployment details for private environments available through its sales channel.
Convert incoming audio to text with Ink-2 for live spoken applications.
Generate streaming spoken audio from text with Sonic-3.6 for conversational interfaces.
Cartesia's managed service provides tools for building and operating voice agents, including provisioned phone numbers and concurrent-call capacity that varies by plan.
Plans include instant voice cloning from Pro upward and professional voice-clone slots on higher tiers. Localizing a voice to an added accent is listed at 225 credits per accent.
Text-to-speech, speech-to-text, and agent concurrency have separate capacity pools, so activity in one service does not use another service's concurrency allowance.
Cartesia lists regional cloud endpoints, on-premise deployment, and on-device deployment for its models and agents; on-premise or virtual private cloud deployment can be discussed with sales.
Use streaming transcription and generated speech to support applications where users speak and receive spoken responses during an interaction.
Build voice agents for calls such as suspicious-transaction verification and step-up authentication, examples described by Cartesia.
Deploy managed agents to handle customer conversations in service operations, with phone-number and concurrent-call allowances depending on the plan.
Evaluate regional cloud processing for in-region workloads, or discuss on-premise and virtual private cloud deployment when infrastructure requirements call for it.
Cartesia offers text-to-speech, speech-to-text, and managed voice-agent services. Sonic-3.6 is its text-to-speech model, Ink-2 is its speech-to-text model, and Managed Agents supports voice-agent workflows.
Plans combine monthly speech-model credits with prepaid agent usage. Agent calls cost $0.06/minute, plus $0.014/minute for Cartesia-provided telephony. The free LLM offer applies to UI-created agents for a limited time; evaluations also have a temporary free offer.
Free costs $0/month and includes 20,000 monthly speech credits and $1 in monthly prepaid agent usage. Pro costs $5/month and adds a commercial-use license and instant voice cloning.
No. Cartesia states that each service has a dedicated concurrency pool, so using text-to-speech does not reduce the available speech-to-text or agent concurrency.
The site lists on-premise and on-device deployment alongside regional cloud endpoints. For on-premise or virtual private cloud deployment, it directs teams to contact sales.
Данные о трафике приведены только для справки.