Cartesia

Cartesia

Cartesia is an AI lab specializing in building real-time, low-latency voice models and agents. Its core products are Sonic (text-to-speech) and Ink (speech-to-text), both ranked #1 for naturalness and accuracy. The platform also offers Managed Agents for building and deploying conversational voice AI, leveraging state space models (SSMs) for efficiency. It is designed for enterprise-grade applications requiring high accuracy, low latency, and automatic speech/turn detection, with deployment options across cloud, on-premise, and on-device environments.

Audio Generation
Visit website
Added on
Sep 21, 2026
Real-time TTSReal-time STTVoice AgentsVoice CloningSSM ArchitectureCloud/On-prem/On-deviceAPI AccessManaged Agents
Product information

Everything worth knowing about Cartesia

The complete picture — from everyday features to the technical detail builders and IT folks dig for.

Core features

What it actually does

Real-time Text-to-Speech
Stream ultra-realistic speech with the Sonic model at sub-90ms latency for live conversation.
Real-time Speech-to-Text
Transcribe audio in real time with the Ink model, featuring native turn detection to know when a speaker starts and finishes.
Voice Agent Platform
Build and deploy low-latency voice agents in minutes. Combines transcription, an LLM, and speech with turn-taking, tools, and telephony handling.
Instant & Professional Voice Cloning
Clone a voice from a short audio clip into 40+ languages and accents. Instant cloning available on Pro plan, professional cloning on higher tiers.
Technical capabilities
State Space Model (SSM) Architecture
Models are built on State Space Models (SSMs), a primitive for low latency, long-context reasoning, and greater efficiency at scale.
Multi-Deployment Options
Same models and agents can be deployed across cloud (regional API endpoints), on-premise, and on-device environments.
Must watch videos
Introducing Line by Cartesia: The Modern Voice Agent Development Platform
5:11
Tutorial5:11
Introducing Line by Cartesia: The Modern Voice Agent Development Platform
Line makes it easy for developers and businesses everywhere to build best-in-class voice agents with code. Learn more | cartesia.ai/agents Read the docs | https://docs.cartesia.ai/line/introduction Check out the SDK | https://github.com/cartesia-ai/line
Cartesia AI July 2026 Changelog & Product Updates
6:06
Tutorial6:06
Cartesia AI July 2026 Changelog & Product Updates
Cartesia AI July 2026 Changelog — Multilingual Voices, Key Term Prompting & More Your monthly rundown of what's new in Cartesia's voice AI platform. This month: multilingual voices go live, Ink-2 gets keyword boosting for tricky pronunciations, and a few housekeeping updates you'll want to know about before October. What's new ✅ Key term prompting (keyword boosting) added to the Ink-2 ASR model ✅ Playground upgrade: transcribe without a mic — use samples or upload your own audio ✅ Multilingual voices launch — 42+ voices, up to 15 languages, native accents ✅ How to clone your own voice into a new language (Hebrew demo) ✅ Locale API: list and set supported languages for any voice ✅ Custom voices are now private by default — how to share them ✅ Admin API expanded to manage org users and invites ✅ Usage dashboard: export credit and call activity to CSV ✅ Deprecation notice: Sonic 2, Sonic Turbo, and Sonic 3 (Oct 2025) sunset October 20, 2026 Timestamps 0:00 – Intro & where to find the full changelog 0:20 – Key term prompting (keyword boosting) lands on Ink-2 ASR 1:00 – Playground update: transcribe without a mic 1:25 – Multilingual voices are here (42+ voices, 15 languages) 2:00 – Locales in action: Gemma in English, Indian-accented English, and Hinglish 2:45 – Cloning your own voice into another language (Hebrew demo) 3:50 – Setting voice locales via the API 4:15 – New default: custom voices are private until shared 4:40 – Admin API: manage org users and invites 5:10 – Usage dashboard: CSV export by model, voice, or API key 5:35 – Deprecations: Sonic 2, Sonic Turbo, Sonic 3 (Oct 2025) — sunset Oct 20, 2026 6:00 – Wrap-up FAQs What is key term prompting on Ink-2? It's keyword boosting / custom vocabulary support on Cartesia's Ink-2 transcription model, so brand names, proper nouns, and unusual spellings get transcribed correctly. Can Cartesia voices speak multiple languages? Yes. As of July 2026, Cartesia voices — including cloned voices — can speak natively in up to 15 languages with locale-specific accents, not just translated text. Are custom cloned voices public by default? No. As of this update, voices you create are private to your organization by default. To use one on a third-party platform, you need to explicitly share it via the API or the "My Voices" tab in the playground. Which Cartesia TTS models are being deprecated? Sonic 2, Sonic Turbo, and the October 2025 version of Sonic 3 are being sunset after October 20, 2026. Cartesia recommends migrating to Sonic 3.5. Can I manage my organization's users via the Cartesia API? Yes. The API now supports listing users, listing and creating invites, and removing users — this requires an admin API key. Resources: ✅ Full July 2026 changelog: https://docs.cartesia.ai/changelog/2026 ✅ Try it in the playground: https://play.cartesia.ai/?utm_source=yt 🔗Ink-2 key term prompting docs: https://docs.cartesia.ai/use-the-api/stt/keyterms?utm_source=yt 🔗 Voice Localization docs: https://docs.cartesia.ai/build-with-cartesia/capability-guides/multilingual-voices#localization?utm_source=yt 🔗 Sonic sunsetting docs: https://docs.cartesia.ai/build-with-cartesia/tts-models/older-models?utm_source=yt Try Cartesia for free at https://play.cartesia.ai/?utm_source=yt 🔗 Questions? Try our Cartesia Subreddit: https://www.reddit.com/r/CartesiaAI/?utm_source=yt 🔗 X/Twitter: https://x.com/cartesia 🔗 LinkedIn: https://www.linkedin.com/company/cartesia-ai
This NEW AI Voice Just DESTROYED ElevenLabs (Cartesia Sonic 3)
15:21
Tutorial15:21
This NEW AI Voice Just DESTROYED ElevenLabs (Cartesia Sonic 3)
Build Your First AI Voice Agent: https://youtu.be/4kzvJccoUow Discover the power of realistic ai voice agent technology with this comparison of ElevenLabs and Cartesia. See how new ai text to speech models stack up against elevenlabs, and how they can be used for text to speech, text to voice, and next generation ai voice agents in real world applications. Check out the newest developments in voice ai, advanced voice cloning, conversational ai agents, and realistic ai voice agent systems designed for customer service, automation, and content creation. 💼 Want to work with me directly? https://arose.ai/booking 🎓 Want to learn more about AI? (for FREE) https://www.skool.com/aa-academy 0:00 – Welcome Back! 1:21 – Inside Cartesia Sonic 3 4:07 – 11Labs vs Cartesia 10:54 – Results and Takeaways

Who uses it

Real use cases, no hype
Customer Service Automation
Deploy 24/7 voice agents for finance, healthcare, etc., with real-time speech understanding.
Real-time Dubbing & Localization
Clone a voice and localize content into 40+ languages and accents on the fly.
Interactive Voice Response (IVR)
Replace rigid IVRs with natural, conversational AI agents for step-up authentication or fraud alerts.
Content Creation & Voiceover
Generate ultra-realistic voiceovers for videos, podcasts, and e-learning at scale.
Transcription & Meeting Notes
Stream highly accurate, real-time speech-to-text with automatic speaker turn detection.
Recruiting & Interviews
Conduct initial screening interviews with AI agents for consistent candidate engagement.
On-device AI Experiences
Deploy fast, private voice AI models directly on user devices for latency-sensitive apps.
What's great
  • #1 ranked speech naturalness & transcription accuracy
  • Ultra-low latency: sub-90ms for first audio byte
  • Unified API for TTS, STT, and voice agents
  • Flexible deployment: cloud, on-premise, on-device
  • Generous free tier for prototyping
Technical strengths
  • Native turn detection in Ink STT model
  • Built on State Space Models (SSMs) for efficiency
Where it falls short
  • Professional voice cloning locked to higher tiers
  • Overage charges can be complex with mixed models
  • Concurrency limits are strict on lower plans
  • Publicly documented models are latest (3.5/2), not all variants
  • Limited concurrent calls on free/Pro tiers
Technical limitations
  • Agent telephony costs are an extra per-minute fee
  • Voice localization costs 225 credits per added accent
For technical folks

The deep-dive specs

FAQ

Includes technical Q&A
Sonic is built on State Space Models (SSMs), not traditional transformers. This enables its signature low latency (~90ms) and high naturalness for real-time conversation.

Traffic Insights

Monthly Visits

—

October 2026

Growth Rate

0%

vs last month

Dominance

0%

in Audio Generation

Top Country

—

Top Countries Breakdown

—

Growth Trend

Stable

Traffic has remained stable.

Honest pricing

No sneaky tiers, no “contact sales”

Here's who each plan is actually for — and where the hidden charges might hit.

Free
Free

Individuals

  • 20K credits / month
Pro
Most picked
$5/mo

Individuals

  • 100K credits / month
Startup
$49/mo

Startups

  • 1.25M credits / month
Scale
$299/mo

Scale

  • 8M credits / month
Enterprise
Custom

Enterprise

  • Custom credits & agent usage
View Full Plans & Pricing on Cartesia

Official pricing, checkout, and plan upgrades are handled directly on cartesia.ai.

Reviews

What the internet actually thinks

Data refreshed weekly. No paid placements.

Trustpilot
Not reviewed yet
G2
G2
Not reviewed yet
C
Capterra
Not reviewed yet
Product Hunt
Not reviewed yet
COMMUNITY COMMENTS

Leave a Comment

Help others make informed decisions. Your honest feedback shapes the community's understanding of this tool.

Tips for commenting

  • •Be specific about features you used
  • •Share real use cases and results
  • •Mention both pros and cons
  • •Keep it honest and constructive
to leave a comment

Your comment is published immediately.

Filter by rating:
Sort by:

No comments yet

Be the first to share your experience with this tool.

Alternatives

Not quite the right fit?

Here's what else is out there.