Cartesia
Also known as: Cartesia Sonic
Real time voice models (Sonic text to speech, Ink speech to text) and Managed Agents, a hosted voice agent platform with tools, a knowledge base, batch calling and call evaluations.
Cartesia builds real time voice models on state space model research and a hosted voice agent platform on top of them. Sonic is its text to speech family, now at Sonic 3.6 across 44 languages with voice cloning, multilingual voices and pronunciation dictionaries. Ink-2 is its streaming speech to text model with built in turn detection and keyterm prompting.
Managed Agents combine Ink-2, an LLM the customer picks from Cartesia's list and Sonic into a realtime voice agent that handles turn taking and interruptions. Agents call webhook tools on the customer's server or client tools in the customer's app, transfer calls to a phone number, and search a knowledge base of documents the customer uploads, which Cartesia indexes on upload. They answer on Cartesia or Twilio numbers, SIP trunks and a WebSocket for web and mobile apps, and call batches dial up to 5,000 recipients, at a scheduled time if one is set. Every configuration change is saved as a version.
Each call keeps a transcript with turn level timestamps, call logs and custom events, and webhooks stream call and turn events including tool calls. Custom LLM judged metrics score completed calls. Enterprise organizations get SSO over SAML or OIDC, dedicated regional deployments in the US, EU, UK, India and Australia, and zero data retention. Cartesia is SOC 2 Type II certified.
Pricing is usage based on monthly plans. The Free, Pro ($5 a month), Startup ($49 a month) and Scale ($299 a month) plans each come with included credits and prepaid agent usage, and Enterprise is sold through sales. Text to speech bills by the character, speech to text by the second and agents at $0.06 a minute.
Vendor details
Canonical URL
https://www.cartesia.ai
Category
Agent infrastructure
Subcategory
Voice infrastructure
Funding status
Independent.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Cartesia offers REST and WebSocket APIs with OpenAPI and AsyncAPI specs, JavaScript/TypeScript and Python SDKs, a CLI and an MCP server. Managed Agents use webhook, client and system tools, a document knowledge base, Cartesia or Twilio numbers and SIP trunks, and call event webhooks. Model plugins bring Cartesia's models to LiveKit, Pipecat, Vapi and other voice stacks.
In practice
Your voice agent feels robotic because of lag. You swap in Cartesia Sonic, which streams first audio in under ninety milliseconds, so responses land before the caller registers a pause.
You need one brand voice across languages. Cartesia clones a voice from ten seconds of audio, and Sonic speaks 44 languages natively.
A regulated deployment cannot send audio to the cloud. Cartesia runs the same Sonic and Ink models on premise or on device, with inference in region for data residency and compliance.
Sources & related URLs
Agentic Index coverage score
10.5 / 14 capabilities · 75%
| Integrations & Tool Calling | Full |
|---|---|
|
Managed Agents take webhook tools, where Cartesia sends the HTTPS request to an endpoint the customer exposes, client tools that run in the customer's app over the WebSocket, and system tools for ending a call, sending DTMF and transferring to a phone number, up to 32 tools per agent, so an agent can look up an order or trigger an action during a call. Plugins for LiveKit, Pipecat and other voice stacks distribute Cartesia's models and are not agent integrations. SourceCartesia, docs.cartesia.ai (Tools, Webhook tools, Integrations)read 2026-09-27 |
|
| Workflow Orchestration | Partial |
|
Cartesia runs a fixed agent loop. A Managed Agent combines Ink-2 listening, the chosen LLM and Sonic speech with turn taking, interruptions and tool calls, and each configuration change is saved as a version. No workflow graph, branching model or hand off between agents is described for Managed Agents, and the older Line SDK is being migrated away from. SourceCartesia, docs.cartesia.ai (Managed Agents, Agent configuration, Migrate from the Line SDK)read 2026-09-27 |
|
| Knowledge Grounding & RAG | Full |
|
Documents uploaded through the Playground or the documents API, singly or in bulk up to 100 a request, are indexed automatically on upload and organized into nested folders with metadata for filtering. An agent's queries search only the folders attached to it, and documents can be updated or deleted through the API. SourceCartesia, docs.cartesia.ai (Knowledge Base, Create Document API)read 2026-09-27 |
|
| Human Oversight & Guardrails | Partial |
|
The transfer system tool hands the caller to a phone number, and turn settings end or check in on silent calls. The configuration's Guardrails are rules written into the agent's instructions, and no approval step holds a tool call for a person. SourceCartesia, docs.cartesia.ai (Agent configuration, System tools)read 2026-09-27 |
|
| Security, Identity & Governance | Full |
|
Enterprise organizations get self serve SSO over SAML or OIDC, configured by an organization admin, and Admin and Member roles, where only admins manage members, invitations and settings. Cartesia states it is SOC 2 Type II certified, a Vanta trust center sits at trust.cartesia.ai, and Zero Data Retention is available on Enterprise. SourceCartesia, docs.cartesia.ai (Set up SSO, Set up an organization) and cartesia.ai/pricingread 2026-09-27 |
|
| Observability & Auditability | Full |
|
Every call has a transcript with turn level timestamps and call logs with the agent's logging statements, custom events can be recorded, runtime logs are retrievable through the API, and webhooks deliver call and turn events carrying each turn's tool calls, variable updates and speech latency, plus a post call analysis. SourceCartesia, docs.cartesia.ai (Observability, Get Call Runtime Logs)read 2026-09-27 |
|
| Memory & State Persistence | Partial |
|
Within a call the agent keeps the conversation, and dynamic variables set at the start or changed by tools are read by the agent each time it responds. Nothing carries into the next call except what the customer passes in as variables, and no memory layer with a scope and lifetime is described. SourceCartesia, docs.cartesia.ai (Agent configuration, Dynamic variables)read 2026-09-27 |
|
| Deployment & Data Residency | Full |
|
Enterprise customers can use dedicated regional deployments in the United States, European Union, United Kingdom, India and Australia that keep inference traffic in region. The same models and agents run cloud, on premise and on device, and on premise and VPC deployment are arranged through sales. SourceCartesia, docs.cartesia.ai (Regional endpoints) and cartesia.ai (home, pricing FAQ)read 2026-09-27 |
|
| Prebuilt Agents, Templates & Packs | Partial |
|
The agents API lists public, Cartesia provided agent templates to help a customer get started, but no template or the job it does is named, and there is no whole agent a customer adopts. The voice library, multilingual voices and pinned model snapshots are model assets, not agents. SourceCartesia, docs.cartesia.ai (List Templates API reference)read 2026-09-27 |
|
| Triggers & Channel Coverage | Full |
|
A call batch stores up to 5,000 recipients, and Cartesia dials them in the background, at a scheduled_at time when one is set, up to a concurrency limit, with retries for failed calls. Agents also take inbound and outbound calls on Cartesia or Twilio numbers and SIP trunks, and stream over a WebSocket from web and mobile apps. SourceCartesia, docs.cartesia.ai (Batch calling, Phone numbers, WebSocket API)read 2026-09-27 |
|
| Model Flexibility & Routing | Full |
|
The customer chooses the model. Each agent's config sets the LLM by ID from the list GET /v1/agents/models returns, with provider, average latency and token prices shown, plus temperature and output limits, and changing it creates a new version so candidates can be compared. Cartesia manages the provider accounts, and speech in and out uses Cartesia's own Ink and Sonic models. SourceCartesia, docs.cartesia.ai (LLMs, Agent configuration)read 2026-09-27 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
A versioned REST API, with 214 reference pages and published OpenAPI and AsyncAPI specs, covers text to speech, speech to text, voices and agents, including agents, tools, knowledge base documents, call batches, metrics and webhooks. Official JavaScript and TypeScript and Python SDKs, a CLI, and an MCP server that runs TTS, STT, voices and pronunciation dictionaries sit beside it. SourceCartesia, docs.cartesia.ai (llms.txt, Client Libraries, MCP)read 2026-09-27 |
|
| Testing, Debugging & Optimization | Partial |
|
Completed calls are scored with built in metrics, such as call success and speech latency, and with custom LLM as a judge metrics the customer writes and assigns to an agent, with results exportable, so output quality can be tracked over time. There is no fixture or dataset testing before production, quality gate or tuning loop. SourceCartesia, docs.cartesia.ai (Metrics, Results)read 2026-09-27 |
|
| Browser & Computer Use | Not documented |
|
The DTMF system tool presses phone keys, which is not browser or desktop control, and the browser examples stream voice into a web page, the product running in a browser. There is no browser, desktop or computer control. SourceCartesia, docs.cartesia.ai (llms.txt, System tools)read 2026-09-27 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Product updates
Managed Agents, Cartesia's hosted voice agent platform, is now available, with its voice models and agent harness designed together so listening, turn taking, tool calls and speech work as one system. It bundles telephony, knowledge bases, backend tool calls and outbound batch campaigns, and agents can switch between English, Spanish, Hindi, French and Japanese mid conversation. Launch pricing is $0.03 a minute, which is 50% off.
Bears on: Agent capability
View sourceCartesia introduced Multilingual Voices, a library of more than 50 voices that support up to 25 languages while retaining the same vocal identity. Customers can keep a voice ID across its supported languages and select a locale in speech generation requests. The offering is available through the voice library and API, with language and accent support varying by voice.
Bears on: Agent capability
View sourceCartesia released Sonic 3.6, an updated text-to-speech model that adapts intonation, pacing, and emotiveness to transcript context without SSML tags. The release adds support for Odia and Urdu, bringing the total to 44 supported languages, and improves transcript following for Hindi and Hinglish.
Bears on: Agent capability
View sourcePricing
Free plan, then Pro $5 a month, Startup $49 a month and Scale $299 a month. Enterprise is custom. Usage is billed in credits and agent minutes.
Credits, counted per character of speech generated and per second of audio transcribed, plus agent minutes.
Included quota
Free includes 20K credits and $1 of agent usage a month. Pro ($5 a month) includes 100K credits and $5 of agent usage. Startup ($49) includes 1.25M credits and $49 of agent usage. Scale ($299) includes 8M credits, $299 of agent usage and priority support. Enterprise has custom credits, agent usage and concurrency. Every plan has unlimited workspace seats, and team invitations start at Startup.
What is public
Plan prices, included credits and agent usage, per credit TTS and STT rates and the per minute agent rate are published. Enterprise is sold through sales.
Billing mechanics
Monthly plan fee with included credits and prepaid agent usage. TTS charges 1 credit per character, STT 1 credit per second, and Managed Agents $0.06 a minute.
Cost watchouts
LLM usage in agents is free only for a limited time, after which LLM tokens bill on top of the agent minute rate. Inviting team members requires Startup or higher. Free plan credits are small for production.
Variable cost rationale
Cartesia meters speech in credits, 1 per character of text to speech and 1 per second of speech to text, and voice agents by the minute at $0.06, plus telephony. The bill grows directly with call volume and the audio generated, and the plan fees are small floors under it. Each plan includes a monthly credit allowance and prepaid agent usage equal to its price, which sets what the fee covers. LLM usage in agents is free for a limited time and then bills on top of the minute rate. Enterprise sets custom credits, agent usage and concurrency.
Overage / add-ons
Usage above included credits bills at 1 credit per character (TTS) and 1 credit per second (STT). Agents bill $0.06 a minute against prepaid agent usage.
Sales call required
Mixed (some tiers require a call)
Free / trial
Free plan with 20K credits and $1 of agent usage a month, without commercial use
Lowest paid plan
Pro at $5 per month, with 100K credits and $5 of agent usage
Commercial notes
Independent. Cartesia holds SOC 2 Type II certification, and Enterprise adds SSO, regional deployments and zero data retention.
Key ambiguities
LLM token prices for agents after the free period are listed per model in the API rather than on the pricing page, and no end date for the free period is published.
Cancellation / refund
Plans can be canceled at any time, with credits retained.
Support SLA / resale
Priority support on Scale. Enterprise adds a shared Slack channel, SSO, DPAs and BAAs.
Related vendors
- AgentOps — Agent observability and debugging platform: open source SDKs trace…
- AgentPhone — AgentPhone gives the AI agents a team builds their own US or…
- Agno — Python agent framework and AgentOS runtime (formerly Phidata) for…
- AIsa — Resource and payment gateway for AI agents, with one key to 110+…
- Albato — Albato is a no code integration platform that sells an embedded,…
- AlphaBitCore — AI control plane for regulated financial firms: one gateway enforces…
Alternatives to Cartesia
The closest documented capability profiles to Cartesia among agent infrastructure platforms tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Sourcegraph11.5 / 14Fuller documented coverage on Workflow Orchestration and Human Oversight & Guardrails
- Paragon10.0 / 14Fuller documented coverage on Workflow Orchestration
- SmythOS12.0 / 14Adds documented Browser & Computer Use
- Thread AI11.0 / 14Fuller documented coverage on Workflow Orchestration and Human Oversight & Guardrails
- xpander.ai12.0 / 14Fuller documented coverage on Workflow Orchestration and Human Oversight & Guardrails
- Kestra12.5 / 14Fuller documented coverage on Workflow Orchestration and Human Oversight & Guardrails
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded