Cartesia

What Cartesia does

Cartesia (Cartesia AI, Inc.) is a San Francisco-based AI infrastructure company founded in September 2023 by Stanford AI Lab researchers Karan Goel, Albert Gu, and Arjun Desai, the inventors of State Space Model (SSM) architectures. The company builds proprietary real-time voice AI foundation models on this SSM primitive — an alternative to the Transformer — delivering 10-100x throughput efficiency and enabling ultra-low-latency inference across cloud, on-premise, and on-device deployments. Its commercial product suite comprises three core offerings: Sonic (text-to-speech, ranked #1 on the Artificial Analysis Speech Arena with sub-90ms latency and 40+ languages), Ink (streaming speech-to-text, ranked #1 on the Artificial Analysis STT leaderboard with native turn detection and semantic endpointing), and Line (a code-first voice agent orchestration platform that combines Sonic and Ink with LLM reasoning into a single production pipeline). The platform supports deployment in cloud, on-premise, on-device, and air-gapped environments, with full enterprise compliance (SOC 2 Type II, HIPAA, GDPR, PCI).

Cartesia monetizes through a hybrid PLG-plus-enterprise model: a self-serve tiered subscription stack (Free $0, Pro $5/mo, Startup $49/mo, Scale $299/mo) accessed via play.cartesia.ai with usage-based overages for voice agent minutes ($0.06/min) and telephony ($0.014/min), combined with custom-volume enterprise contracts for large customers including ServiceNow, Retell AI, Replicant, Forethought, Cresta, Together AI, and Maven AGI. Distribution is materially amplified by OEM-style embedding across voice agent platforms where Cartesia is the default or primary voice provider — Vapi, LiveKit, Daily, Quora/Poe, Rasa, Tavus, Together AI, and Cerebrium — exposing Cartesia voices to tens of millions of monthly calls.

Cartesia has raised approximately $100M in cumulative funding across a $27M seed (December 2024, Index Ventures) and $64M Series A (March 2025, Kleiner Perkins), with strategic capital from NVIDIA, Ericsson, and Databricks. The company has ~100 employees (up from 20 at founding), with a ~10-person team dedicated to India operations launched in January 2026 from a Bengaluru office supported by an India-resident data center. Its primary customer segments are enterprise voice agent builders (ServiceNow, Retell, Forethought, Replicant), regulated financial services and healthcare buyers (Rasa, Blue Machines AI partnerships, Fundamento for Indian BFSI), AI infrastructure and platform providers (Vapi, LiveKit, Together AI), and a long tail of individual developers and startups on the free and Pro tiers.

Name

Cartesia

Legal name

Cartesia AI, Inc.

Website

https://cartesia.ai

Company type

Private

Founded year

2023

Operating status

Operating

Headcount range

51–100 employees

Short description

Cartesia provides real-time multimodal voice AI infrastructure, offering proprietary text-to-speech (Sonic), speech-to-text (Ink), and voice agent orchestration (Line) products built on State Space Model architecture for enterprise platforms and developer customers globally.

Ownership category

akta.pro rank

Product category

Real-Time Voice AI Platform

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518), Custom Computer Programming Services (541511)

SIC

Services-Computer Programming, Data Processing, Etc. (7370), Services-Prepackaged Software (7372)

akta.pro primary industry

Voice Assistants & Voice UI (VUI) (HDAAAFAD)

akta.pro secondary industries

Contact Center Speech AI (agent assist, IVR automation) (HDAAAFAL), Model Hosting, Serving & Inference Platforms (HDAAACAB)

Keywords

• Real-time voice AI
• Speech-to-text API
• Text-to-speech API
• Voice agent platform
• State space models

Headquarters

HQ city

San Francisco

HQ country

United States

HQ region

North America

Offices2 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Infrastructure, Marketing or Sales, Operations

Revenue model

1. Self-serve SaaS subscriptions: Tiered monthly subscriptions: Free ($0), Pro ($5/mo promotional $4), Startup ($49/mo promotional $37), Scale ($299/mo promotional $224). Plans include monthly credit allocations for Sonic TTS and Ink STT, agent slots, concurrent call limits, and progressively richer features (commercial license, instant/pro voice cloning, organizations, priority support, high concurrency). June 2026 promotion offers 25% off 12 monthly cycles for new sign-ups to Pro/Startup/Scale.
2. Usage-based consumption overage: Pay-as-you-go overage on top of subscription credits for additional TTS minutes, STT hours, and voice agent call minutes. Voice changer at 15 credits/second; voice localization 225 credits one-time; voice agent call duration $0.06/minute; Cartesia-provided telephony $0.014/minute.
3. Enterprise / Custom contracts: Custom credits and agent usage, volume pricing, custom concurrency limits, DPAs/BAAs for compliance, shared Slack channel, SSO, and security questionnaires for large enterprise customers (e.g., ServiceNow).
4. Freemium entry tier: Free tier with 20K credits/month and $1 prepaid agent minutes/month for prototyping, no commercial use license, used to drive self-serve funnel conversion.

Pricing tiers

Go-to-market motion6 records

Distribution channels7 records

Marketing channels9 records

Core offering

Cartesia provides a unified real-time voice AI platform built on proprietary State Space Model (SSM) architecture, delivering Sonic text-to-speech, Ink speech-to-text, and Line voice agent orchestration through a single API. The platform serves developers and enterprises building production voice agents for customer support, financial services, healthcare, and localization, with deployment options spanning cloud, on-premise, and on-device.

Product overview

Cartesia offers a unified real-time voice AI platform built on proprietary State Space Model (SSM) architecture, architected as three core products that work together end-to-end: Sonic (the flagship text-to-speech model, ranked #1 for naturalness with sub-90ms latency across 40+ languages), Ink (the streaming speech-to-text model, ranked #1 on accuracy with native turn detection), and Line (the code-first voice agent development platform that combines Sonic and Ink with LLM orchestration into a production-ready conversational AI pipeline). The platform is augmented by capabilities including Professional Voice Cloning for brand voices, Voice Changer for audio transformation, and Narrations for long-form audio content creation. Cartesia supports flexible deployment across cloud, on-premise, and on-device environments, with enterprise-grade security certifications (SOC 2 Type 2, HIPAA, GDPR, PCI) and 40+ language support — serving customers like ServiceNow, Retell, Vapi, Rasa, Tavus, Tencent Cloud, and Fundamento across customer support, financial services, healthcare, gaming, and conversational AI use cases.

Differentiator

Problem solved

Functional benefit

Brands

• Sonic: Cartesia's real-time text-to-speech model family (including Sonic-3.5), supporting 40+ languages with sub-90ms latency.
• Ink
• Line

Products and services

• Sonic (Sonic-3.5) Sonic is Cartesia's flagship real-time text-to-speech model, ranked #1 on the Artificial Analysis Speech Arena. Built on State Space Model architecture, it delivers sub-90ms latency with native multilingual support across 40+ languages and regional variants, emotion/laughter synthesis, instant voice cloning from 10 seconds of audio, voice localization, custom pronunciation dictionaries, and voice changer. Designed for developers and enterprises building real-time voice agents, customer support, and conversational AI applications.
• Ink (Ink-2) Ink is Cartesia's streaming speech-to-text model, ranked #1 on the Artificial Analysis Speech-to-Text leaderboard. It delivers the lowest Word Error Rate of any streaming STT model with 100ms time-to-final-transcript latency and native turn detection (turn.start, turn.end, turn.eager_end). Features include semantic endpointing and structured data handling for phone numbers, dates, currencies, and UUIDs.
• Line Line is Cartesia's code-first voice agent development platform that combines Ink STT and Sonic TTS into a unified real-time pipeline for production-ready enterprise voice agents. Supports custom LLM integration (GPT-5.4, Gemini 3.5 Flash), parallel background tasks, real-time system actions, built-in evaluations, SIP trunking, and observability. Ships via Python SDK and other SDKs, deployable in cloud, on-premise, or on-device, with enterprise features including SSO, custom concurrency limits, DPAs, BAAs, and security questionnaires.
• Narrations Narrations is a Cartesia platform introduced in March 2025 that enables creators to transform written content into polished long-form audio productions with precision control and editing efficiency for podcasts, audiobooks, and audio content workflows.

Named customers2 records

Segments9 records

Ideal customer profiles6 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration17 records

AI capability9 records

Feature7 records

Partnerships

Ten partnerships are on record, tiered minor / ecosystem partner, core / india flagship gtm partner, core / asia-pacific flagship partner, flagship / enterprise flagship customer and integration partner, flagship / default voice partner, flagship / enterprise flagship customer and investor, core / embedded distribution partner, core / default voice provider and core / strategic integration partner.

• Runwayminor / ecosystem partnerGTM or Marketing Partner · 31 March 2026Cartesia is one of six founding cohort startups in Runway's $10M Builders program, receiving 500,000 free API credits, access to the Characters real-time video agent API, technical mentorship, and enterprise customer introductions. Runway's fund thesis focuses on technical teams pushing AI architecture frontiers.
• Blue Machines AIcore / india flagship gtm partnerStrategic or Co-development Partner · 20 February 2026Strategic partnership to deliver India-first conversational AI solutions for Indian enterprises with initial focus on regulated BFSI and healthcare sectors. Enables India-resident data processing and multilingual support across Tamil, Telugu, Kannada, Marathi, and other Indian languages. Co-developed jointly tested deployment architecture with attention to latency, audio handling, interruption control, and operational governance.
• Tencent Cloudcore / asia-pacific flagship partnerStrategic or Co-development Partner · 25 November 2025Strategic partnership to advance real-time Voice AI experiences by combining Tencent's global communication infrastructure with Cartesia's Sonic 3 model. Targets AI voice applications in Southeast Asia and Africa; supports over 40 languages with human-like qualities.
• ServiceNowflagship / enterprise flagship customer and integration partnerStrategic or Co-development Partner · 25 October 2025ServiceNow integrated Cartesia's Sonic TTS, Ink STT, and Line voice agent orchestration with its AI Voice Agents (part of the AI Experience platform). The integration supports fully air-gapped on-prem deployment and SOC 2 Type II, HIPAA, GDPR, and PCI compliance. Ravi Krishnamurthy (VP of Product, AI Platform and Responsible AI at ServiceNow) is quoted publicly on the partnership.
• Rasaflagship / default voice partnerStrategic or Co-development PartnerCartesia selected as the default voice provider for Rasa's enterprise conversational AI platform, combining Rasa's AI assistant capabilities with Sonic TTS (90ms TTFA, 200+ ultra-realistic voices, SOC2 and HIPAA compliance). Targets American Express, Adobe, T-Mobile and reports 181% ROI and 60% containment rate for enterprise customers.
• Ericssonflagship / enterprise flagship customer and investorStrategic or Co-development PartnerEricsson uses Cartesia TTS to power its enterprise avatar assistant Corina for customer service, alongside Tavus (avatar technology) and Pipecat (virtual room platform). Cartesia is also an Ericsson investment. Ericsson demonstrated three AI voice services including deepfake voice detection and real-time scam detection.
• Together AIcore / embedded distribution partnerChannel Partner/ Reseller/ DistributorCartesia Sonic TTS available natively on Together AI's AI acceleration cloud, serving millions of audio minutes on Together's Voice Platform as a dedicated model partner for enterprise voice AI workloads.
• LiveKitcore / embedded distribution partnerTechnology or IntegrationLiveKit partnered with Cartesia because Sonic, built on State Space Models, aligned perfectly with LiveKit's vision for next-generation AI agents. Sonic is integrated into LiveKit's platform for real-time AI agent voice interactions.
• Vapicore / default voice providerTechnology or IntegrationCartesia is Vapi's default voice provider for voice agents. Vapi enables anyone to build, test, and deploy high quality conversational voice agents in minutes.
• Tavuscore / strategic integration partnerTechnology or IntegrationTavus chose Cartesia's Sonic model as the cornerstone of their voice generation system for the world's fastest conversational video interface. Sonic's State Space Model capabilities aligned with Tavus's need for high-quality, low-latency voice generation.

Scale indicators14 records

Recent moves7 records

Expansion highlights8 records

Direct peers

• ElevenLabs: ElevenLabs is the leading independent TTS and voice AI platform, offering speech synthesis, voice cloning, dubbing, and conversational agents across dozens of languages. It directly competes with Cartesia's Sonic TTS and voice agent capabilities for both developers and enterprise customers.
• Deepgram: Deepgram provides enterprise-grade speech-to-text APIs with low-latency streaming transcription, voice agent infrastructure, and medical/contact-center models. It directly competes with Cartesia's Ink STT and Line voice agent platform on accuracy, latency, and price-performance.
• AssemblyAI: AssemblyAI offers speech-to-text, speech understanding (speaker diarization, sentiment, topic detection), and LLM-powered audio intelligence APIs. It competes directly with Cartesia's Ink STT and adjacent transcription/intelligence workloads for developer and enterprise buyers.
• Speechmatics: Speechmatics provides accurate, multilingual speech-to-text with emphasis on enterprise and contact-center deployments and on-prem options. It overlaps directly with Cartesia's Ink STT product and the regulated-enterprise motion Cartesia is pursuing.

Broad incumbents

• OpenAI: OpenAI ships Realtime API for speech-to-speech conversation, TTS, and Whisper-based STT as part of its broader GPT platform. It is a broad incumbent that competes with Cartesia's Sonic, Ink, and Line offerings, typically bundled with OpenAI's foundation-model stack.
• Google Cloud Speech: Google Cloud offers Cloud Text-to-Speech and Speech-to-Text APIs plus Chirp and Gemini-powered speech models as part of its broader AI/cloud portfolio. It competes broadly with Cartesia in TTS, STT, and conversational AI for enterprise and regulated workloads.
• Microsoft Azure AI Speech: Azure AI Speech delivers TTS, STT, and real-time voice translation as part of Microsoft's Azure AI stack, often bundled with Teams, Copilot, and enterprise agreements. It is a broad incumbent competitor in enterprise voice AI.
• Amazon Web Services (Polly / Transcribe): AWS Polly (TTS), Amazon Transcribe (STT), and Bedrock-powered voice agents compete with Cartesia as part of the dominant cloud platform's full-stack AI portfolio, particularly for regulated enterprise and government workloads.

Emerging players

• Hume AI: Hume AI focuses on emotionally intelligent voice interfaces with its Empathic Voice Interface and expression-measurement models. It is an emerging player overlapping with Cartesia's Sonic emotional TTS and the broader real-time conversational voice category.
• PlayAI (PlayHT): PlayAI (PlayHT) provides realistic TTS, voice cloning, and conversational voice agents aimed at developers and enterprise content use cases. It is an emerging player that directly overlaps with Cartesia's Sonic TTS and agent offerings.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat7 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/cartesia_ai
• GitHubgithub.com/cartesia-ai
• YouTubeyoutube.com/@cartesia_ai
• Facebookfacebook.com/cartesiaai
• LinkedInlinkedin.com/company/cartesia-ai
• Instagraminstagram.com/cartesia_ai

Compliance4 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles4 records

Funding overview

Funding rounds4 records

Investors13 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Cartesia