Showing the first 18000 characters. Load the complete file to read the remainder.
# Cartesia Docs
- [API (215 pages)](https://docs.cartesia.ai/_llms/en/api.txt): Documentation for API.
## Models
### Get started
#### Overview
- [Welcome to Cartesia](https://docs.cartesia.ai/get-started/overview.md): Cartesia is an AI lab building real-time voice models and agents.
#### Start building
- [Realtime Text-to-Speech Quickstart](https://docs.cartesia.ai/get-started/realtime-text-to-speech-quickstart.md): Stream text to Cartesia over a WebSocket and receive audio in real time.
- [Realtime Speech-to-Text Quickstart](https://docs.cartesia.ai/get-started/realtime-speech-to-text-quickstart.md): Stream audio to Ink 2 and receive transcripts in real time.
- [Authenticate your applications](https://docs.cartesia.ai/get-started/authenticate-your-client-applications.md): Use access tokens in browsers and API keys on servers
- [Pricing](https://docs.cartesia.ai/pricing.md): How Cartesia charges for usage
#### Developers
- [Client Libraries](https://docs.cartesia.ai/tools/client-libraries.md): Use the official JavaScript/TypeScript and Python SDKs to call the Cartesia API.
- [AI Agent Guide](https://docs.cartesia.ai/tools/ai/agent-guide.md): Set up your agents with Cartesia skills, MCP, and docs
- [Agent Skills](https://docs.cartesia.ai/tools/ai/agent-skills.md): Teach your agents how to build with Cartesia
- [MCP](https://docs.cartesia.ai/tools/ai/mcp.md): Connect Cartesia MCP in Cursor or Claude Code to run TTS, STT, voices, and pronunciation dictionaries from your agent.
- [LLMs documentation files](https://docs.cartesia.ai/tools/ai/llms-txt.md): Machine-readable index files for assistants and tooling that ingest Cartesia documentation.
#### Integrations
- [Integrations](https://docs.cartesia.ai/integrations/overview.md)
- [Partner Integrations](https://docs.cartesia.ai/integrations/partners.md): Official Cartesia integrations on third-party platforms
- [LiveKit](https://docs.cartesia.ai/integrations/live-kit.md): Realtime rooms and agents via the Cartesia plugin or LiveKit Inference
- [Pipecat](https://docs.cartesia.ai/integrations/pipecat.md): Voice and multi-modal agents with official Cartesia TTS/STT services
##### Community
- [Community Integrations](https://docs.cartesia.ai/integrations/community/index.md): Integrations built and maintained by third-party developers
- [Anam + Cartesia](https://docs.cartesia.ai/integrations/community/anam-cartesia.md): Use a Cartesia Sonic TTS voice in an Anam interactive avatar session
- [Genesys + Cartesia Sonic](https://docs.cartesia.ai/integrations/community/genesys-tts-connector.md): Configure the Genesys TTS Connector integration to synthesize speech with Cartesia
- [LiveAvatar + Cartesia](https://docs.cartesia.ai/integrations/community/liveavatar-cartesia.md): Drive a LiveAvatar (by HeyGen) interactive avatar with a Cartesia Sonic TTS voice
- [TrueFoundry + Cartesia](https://docs.cartesia.ai/integrations/community/truefoundry.md): Route Cartesia TTS and STT through TrueFoundry's AI Gateway
- [Rasa](https://docs.cartesia.ai/integrations/rasa.md): Voice assistants with Cartesia as the TTS backend
- [Tencent RTC](https://docs.cartesia.ai/integrations/tencent-rtc.md): Realtime media with Cartesia for conversational AI workloads
- [Thoughtly](https://docs.cartesia.ai/integrations/thoughtly.md): No-code phone agents
- [Twilio](https://docs.cartesia.ai/integrations/twilio.md): Call a phone with speech synthesized by Sonic
- [Bandwidth + Cartesia](https://docs.cartesia.ai/integrations/community/bandwidth-cartesia.md): Run a two-way voice agent on a Bandwidth phone call with Cartesia Ink 2 speech-to-text and Sonic text-to-speech
- [Vision Agents by Stream](https://docs.cartesia.ai/integrations/vision-agents-by-stream.md): Agent framework with a Cartesia TTS plugin
### Text-to-Speech
#### Sonic
- [Sonic 3.6](https://docs.cartesia.ai/build-with-cartesia/tts-models/latest.md)
- [Build with Sonic](https://docs.cartesia.ai/build-with-cartesia/capability-guides/tts.md)
#### Choose a voice
- [Choosing a Voice](https://docs.cartesia.ai/build-with-cartesia/capability-guides/choosing-a-voice.md): How to pick the best voice for your Voice Agents
- [Multilingual Voices](https://docs.cartesia.ai/build-with-cartesia/capability-guides/multilingual-voices.md): Reach a global audience with one voice that speaks multiple languages and accents.
- [Instant Voice Clone](https://docs.cartesia.ai/build-with-cartesia/capability-guides/clone-voices.md): Create a custom voice in seconds from a short recording
- [Pro Voice Clone](https://docs.cartesia.ai/build-with-cartesia/capability-guides/clone-voices-pro.md): Create a near-exact voice replica, trained on 30+ minutes of your audio
#### Control the voice
- [Prompting tips](https://docs.cartesia.ai/build-with-cartesia/capability-guides/prompting-tips.md): Get natural-sounding output from Sonic with minimal prompt engineering.
- [Custom Pronunciations](https://docs.cartesia.ai/build-with-cartesia/capability-guides/custom-pronunciations.md): Specify custom pronunciations for words that are hard to get right, like proper nouns or domain-specific terms.
- [Text Normalization](https://docs.cartesia.ai/build-with-cartesia/capability-guides/text-normalization.md): Control how written text becomes spoken text, and debugging tips.
- [Volume, Speed, and Emotion](https://docs.cartesia.ai/build-with-cartesia/capability-guides/volume-speed-emotion.md): Control the speed, volume, and emotion of generated speech.
- [SSML Tags](https://docs.cartesia.ai/build-with-cartesia/capability-guides/ssml-tags.md): Laughter, pauses, and mid-transcript controls
- [Advanced capabilities](https://docs.cartesia.ai/build-with-cartesia/capability-guides/advanced-capabilities.md): Use Hinglish code-switching and text normalizers to control how Sonic speaks specialized content.
#### Streaming
- [Compare TTS Endpoints](https://docs.cartesia.ai/use-the-api/compare-tts-endpoints.md): How bytes, SSE, and WebSocket differ for text-to-speech, and when to use each.
- [Stream Inputs using Continuations](https://docs.cartesia.ai/build-with-cartesia/capability-guides/stream-inputs-using-continuations.md): Learn how to stream input text to Sonic TTS
- [Contexts and Continuations](https://docs.cartesia.ai/use-the-api/tts-websocket/contexts.md): Generating audio from partial transcripts
- [Context Flushing and Flush IDs](https://docs.cartesia.ai/use-the-api/tts-websocket/context-flushing-and-flush-i-ds.md): Separating a continuation into multiple parts
- [Buffering](https://docs.cartesia.ai/use-the-api/tts-websocket/buffering.md): Control how text is buffered before speech generation to balance prosody and latency
#### Audio output
- [Output Format](https://docs.cartesia.ai/build-with-cartesia/capability-guides/tts-output-audio-format.md): How to find the right parameters for output audio
- [Caching Audio for Stock Responses](https://docs.cartesia.ai/build-with-cartesia/capability-guides/tts-caching.md): Pre-generate stock TTS phrases as raw PCM, then interleave them with live WebSocket audio to cut latency and credits
#### Older models
- [Sonic 3.5](https://docs.cartesia.ai/build-with-cartesia/tts-models/sonic-3-5.md)
- [Migrating from Sonic 3 to Sonic 3.5](https://docs.cartesia.ai/build-with-cartesia/tts-models/sonic-3-to-sonic-3-5.md): What's new in Sonic 3.5 and what to know before you switch.
- [Older Models](https://docs.cartesia.ai/build-with-cartesia/tts-models/older-models.md)
- [Deprecated Models](https://docs.cartesia.ai/build-with-cartesia/tts-models/api-changes.md)
### Speech-to-Text
- [Ink 2](https://docs.cartesia.ai/build-with-cartesia/stt/latest.md)
- [Older Models](https://docs.cartesia.ai/build-with-cartesia/stt/older-models.md)
#### Guides
- [Build with Ink](https://docs.cartesia.ai/build-with-cartesia/stt/guides.md)
- [Turn Detection](https://docs.cartesia.ai/use-the-api/stt/turns.md): Detect user turns automatically, no separate VAD required
- [Keyterm Prompting](https://docs.cartesia.ai/use-the-api/stt/keyterms.md): Improve transcription accuracy for specific words and phrases
- [Compare STT Endpoints](https://docs.cartesia.ai/use-the-api/stt/compare-endpoints.md): There are 3 ways to integrate with Cartesia's speech-to-text API
- [Audio Input](https://docs.cartesia.ai/build-with-cartesia/stt/audio-input.md): How to find the right encoding and sample rate for realtime audio
- [Migration Guides](https://docs.cartesia.ai/use-the-api/stt/migrations.md): The differences between Ink and other models
##### Troubleshooting
- [Troubleshooting](https://docs.cartesia.ai/use-the-api/stt/troubleshooting/index.md): Transcription errors, high latency, server errors
- [Realtime STT (Auto)](https://docs.cartesia.ai/use-the-api/stt/troubleshooting/auto.md): Transcript errors, high latency, and server errors for `/stt/turns/websocket`
- [Realtime STT (Manual)](https://docs.cartesia.ai/use-the-api/stt/troubleshooting/manual.md): Transcript errors, high latency, and server errors for `/stt/websocket`
## Agents
### Get started
- [Managed Agents](https://docs.cartesia.ai/agents/introduction.md): Create and run realtime voice agents with Cartesia.
### Agent management
- [Agent configuration](https://docs.cartesia.ai/agents/configuration.md): Configure how an agent behaves, listens, speaks, and acts.
- [Dynamic variables](https://docs.cartesia.ai/agents/dynamic-variables.md): Personalize each call with customer data and values from your tools.
- [Managed LLMs](https://docs.cartesia.ai/agents/models.md): Choose an LLM to generate agent responses.
### Tools
- [Tools](https://docs.cartesia.ai/agents/tools.md): Give an agent actions it can invoke during a call.
- [System tools](https://docs.cartesia.ai/agents/system-tools.md): Configure built-in actions for calls and telephony.
- [Webhook tools](https://docs.cartesia.ai/agents/webhook-tools.md): Let an agent call an HTTPS endpoint on your server.
- [Client tools](https://docs.cartesia.ai/agents/client-tools.md): Run functions in your application over the WebSocket API.
### Audio integration
- [Audio integration](https://docs.cartesia.ai/line/integrations/overview.md): Connect an application or phone number to an agent.
- [WebSocket API](https://docs.cartesia.ai/line/integrations/websocket-api.md): Stream audio and events between your application and an agent.
### Telephony
- [Phone numbers](https://docs.cartesia.ai/line/integrations/telephony/phone-numbers.md): Connect phone numbers to your agents for inbound and outbound calls.
- [Cartesia numbers](https://docs.cartesia.ai/line/integrations/telephony/cartesia-numbers.md): Provision and manage Cartesia phone numbers for your agents.
- [Twilio integration](https://docs.cartesia.ai/line/integrations/telephony/twilio/integration.md): Import phone numbers from your Twilio account.
- [Regional routing](https://docs.cartesia.ai/line/integrations/telephony/twilio/regional-routing.md): How Twilio regions affect call routing and credential scoping.
- [SIP trunking](https://docs.cartesia.ai/line/integrations/telephony/sip-trunking.md): Connect a SIP trunk to use your own carrier with Cartesia agents.
- [Outbound calling](https://docs.cartesia.ai/line/integrations/telephony/outbound-dialing.md)
- [Batch calling](https://docs.cartesia.ai/line/integrations/telephony/batch-calling.md)
### Evaluations
- [Metrics](https://docs.cartesia.ai/line/evaluations/metrics.md): Measure agent performance on completed calls.
- [Results](https://docs.cartesia.ai/line/evaluations/results.md): Review and export metric results for completed calls.
### Migration
- [Migrate from the Line SDK](https://docs.cartesia.ai/agents/migrate-from-line-sdk.md): Move a Line SDK agent to Managed Agents, feature by feature.
## Examples
### Text-to-Speech (WebSocket)
- [WebSocket Basic](https://docs.cartesia.ai/examples/tts-websocket-basic.md): Basic WebSocket usage with websocket_connect() context manager.
- [WebSocket with Input Continuations](https://docs.cartesia.ai/examples/tts-websocket-continuations.md): Streaming a transcript split into multiple parts, using continuations.
- [WebSocket Flushing](https://docs.cartesia.ai/examples/tts-websocket-flushing.md): Demonstrates manual flushing to separate audio from different transcripts.
- [WebSocket Emotion Control](https://docs.cartesia.ai/examples/tts-websocket-emotion.md): Demonstrates changing emotion mid-stream using generation_config.
- [WebSocket Speed Control](https://docs.cartesia.ai/examples/tts-websocket-speed.md): Demonstrates changing speed mid-stream using generation_config.
- [WebSocket Concurrent Contexts](https://docs.cartesia.ai/examples/tts-websocket-concurrent-contexts.md): Two contexts on one connection, each using ctx.receive() to get their own audio.
- [WebSocket Response Handling](https://docs.cartesia.ai/examples/tts-websocket-response-handling.md): WebSocket response type handling.
### Text-to-Speech (SSE)
- [SSE Streaming](https://docs.cartesia.ai/examples/tts-sse-basic.md): Basic SSE streaming.
- [SSE with Word Timestamps](https://docs.cartesia.ai/examples/tts-sse-with-timestamps.md): SSE streaming with word timestamps.
- [SSE with Phoneme Timestamps](https://docs.cartesia.ai/examples/tts-sse-with-phoneme-timestamps.md): SSE streaming with phoneme timestamps.
### Text-to-Speech (Batch)
- [Generate to File](https://docs.cartesia.ai/examples/tts-generate-to-file.md): Use generate() and write_to_file() to write a wav file.
- [Create Infill Audio](https://docs.cartesia.ai/examples/infill-create.md): Create infill audio between two clips.
### Speech-to-Text
- [Realtime STT (Auto)](https://docs.cartesia.ai/examples/stt-auto-finalize-websocket.md): Realtime transcription with native turn detection — recommended for voice agents.
- [Realtime STT (Manual)](https://docs.cartesia.ai/examples/stt-manual-finalize-websocket.md): Realtime transcription where you control the user turn. Recommended for push-to-talk apps.
- [Batch STT](https://docs.cartesia.ai/examples/stt-transcribe.md): Transcribe pre-recorded audio files.
### Voices
- [List Voices](https://docs.cartesia.ai/examples/voices-list.md): List voices with pagination.
- [Get a Voice](https://docs.cartesia.ai/examples/voices-get.md): Get a specific voice.
- [Clone a Voice](https://docs.cartesia.ai/examples/voices-clone.md): Clone a voice from an audio clip.
- [Update a Voice](https://docs.cartesia.ai/examples/voices-update.md): Update a voice.
- [Delete a Voice](https://docs.cartesia.ai/examples/voices-delete.md): Delete a voice.
### Browsers
- [Play Audio in Browser](https://docs.cartesia.ai/examples/tts-play-audio.md): Generate a wav and play it using an <audio> element.
- [Download Audio File](https://docs.cartesia.ai/examples/tts-download-file.md): Generate audio and trigger a file download in the browser.
- [WebSocket Stream to Web Audio](https://docs.cartesia.ai/examples/tts-websocket-stream-audio.md): Stream audio from a WebSocket and play it in real-time with Web Audio API.
- [WebSocket Low-Latency Playback](https://docs.cartesia.ai/examples/tts-websocket-low-latency.md): Play audio chunks as they arrive for lowest latency.
### Integrations
- [Next.js Full Example](https://docs.cartesia.ai/examples/nextjs.md): A complete Next.js application with batch TTS, HTTP streaming, and STT transcription.
### Best Practices
- [Error Handling](https://docs.cartesia.ai/examples/error-handling.md): Example of error handling with SDK exceptions.
## Enterprise
### Overview
- [Set up SSO](https://docs.cartesia.ai/enterprise/set-up-sso.md): Configure Single Sign-On (SSO) for your Enterprise organization directly from the Cartesia Playground.
- [Set up an organization](https://docs.cartesia.ai/enterprise/set-up-an-organization.md)
- [Zero Data Retention](https://docs.cartesia.ai/enterprise/zero-data-retention.md): Zero Data Retention (ZDR) is available to Cartesia Enterprise customers for our Text-to-Speech (TTS) and Speech-to-Text (STT) APIs.
- [Regional endpoints](https://docs.cartesia.ai/enterprise/regional-endpoints.md): Dedicated regional deployments are available to Cartesia Enterprise customers to support data sovereignty requirements.
## Changelog
### By year
- [Changelog 2026](https://docs.cartesia.ai/changelog/2026.md): Product, API, and platform changes for 2026
- [Changelog 2025](https://docs.cartesia.ai/changelog/2025.md): Product, API, and platform changes for 2025
- [Changelog 2024](https://docs.cartesia.ai/changelog/2024.md): Product, API, and platform changes for 2024
## Other
- [Delete Agent](https://docs.cartesia.ai/2026-03-01/api-reference/agents/agents/delete.md)
- [Get Agent](https://docs.cartesia.ai/2026-03-01/api-reference/agents/agents/get.md): Returns the details of a specific agent
- [List Agents](https://docs.cartesia.ai/2026-03-01/api-reference/agents/agents/list.md): Lists all agents associated with your account
- [List Templates](https://docs.cartesia.ai/2026-03-01/api-reference/agents/agents/templates.md): List of public, Cartesia-provided agent templates to help you get started
- [Update Agent](https://docs.cartesia.ai/2026-03-01/api-reference/agents/agents/update.md)
- [Get Call Runtime Logs](https://docs.cartesia.ai/2026-03-01/api-reference/agents/calls/get-call-logs.md)
- [Get Deployment](https://docs.cartesia.ai/2026-03-01/api-reference/agents/deployments/get-deployment.md): Get a deployment by its ID.
- [List Deployments](https://docs.cartesia.ai/2026-03-01/api-reference/agents/deployments/list-deployments.md): List of all deployments associated with an agent.
- [Create Document](https://docs.cartesia.ai/2026-03-01/api-reference/agents/documents/create.md): Upload a document into a [knowledge base](/line/knowledge-base) folder
- [Bulk Create Documents](https://docs.cartesia.ai/2026-03-01/api-reference/agents/documents/create-bulk.md): Upload up to 100 [knowledge base](/line/knowledge-base) documents in a single request
- [Delete Document](https://docs.cartesia.ai/2026-03
Measured from the saved file
Linked document
UTF-8 text size
30.2 KB30947 bytes
Token count
7684o200k_base encoding
Markdown links
244207 distinct URLs
Fenced code blocks
0330 lines in the file
Size measures the captured UTF-8 text, not compressed transfer size. Tokens are counted with a named tokenizer; other models can count differently. Neither measure establishes context compatibility or answer quality.
Project H1
Cartesia Docs
Context architecture
How the content is divided
9 H2 sections organize 244 parsed Markdown links. 191 links point to .md or .markdown paths; 1 link occurrences use another hostname. Linked destinations have not been recursively validated.
H2 section
Links
UTF-8 bytes
Tokens
ModelsLine 5
59
8909
2191
AgentsLine 112
20
2727
648
ExamplesLine 155
26
3715
866
EnterpriseLine 207
4
696
159
ChangelogLine 216
3
339
104
OtherLine 224
118
13448
3388
OpenAPI SpecsLine 306
4
113
51
AsyncAPI SpecsLine 313
5
314
104
IndexesLine 323
4
467
119
Link counts exclude parsed fenced and inline code and resolve relative destinations against the final endpoint. Section sizes cover the text below each H2 until the next H2.
Embedded material
Code, installation and API examples
Fenced blocks
0 closed
Declared languages
No fenced blocks detected
Package-install commands
0 detected in fenced or inline code
Authentication examples
0 fenced lines match API-key, Authorization, Bearer, OAuth or access-token patterns
HTTP request examples
0 fenced lines match request-method patterns
Authentication links
0 links have authentication-related labels
Embedded example text and links to further documentation are counted separately. A code block does not establish that a sample runs or that an authentication flow succeeds.
Publisher prioritization
The Optional section
No H2 section named “Optional” was detected in this file.
The community proposal uses Optional by convention for secondary links that can be skipped for a shorter context. This label does not establish that everything outside it is critical, or that the material inside it is deprecated.
Machine-readable routes
MCP references and configuration text
MCP-related text was detected in the captured file. 1 links have MCP-related labels or paths. 0 fenced lines match mcpServers, mcp_servers or an mcp add command.
A verified text response establishes a companion at the checked URL. It does not prove the bundle contains all product documentation, that generation is automated, or that it fits an assistant’s context window.
Change evidence
What can be said about freshness
This capture establishes a baseline. Update frequency cannot be determined from one observation.
Linked release and migration material
5 links have release, changelog, deprecation or migration terms in their labels or paths. These provide places to investigate product changes separately.
File changes can reveal documentation edits or navigation changes. They do not by themselves measure product release velocity, accuracy, agent adoption or AI visibility.
Counts use a deterministic parser, not a spec-compliance grade. Unusual Markdown and indented code can fall outside its recognized patterns. See the llms.txt proposal and tokenizer implementation.