# Cartesia Sonic

> Ultra-low-latency streaming text-to-speech and instant voice cloning model.

Sonic is an AI text-to-speech model built on State Space Model (SSM) architecture, designed for real-time applications and conversational voice agents. It generates expressive speech with sub-90ms latency across 44 languages and can clone voices from just a few seconds of clean reference audio. It also supports inline transcript controls like custom pronunciations and expressive cues like laughter.

Website: https://cartesia.ai/sonic

## What it is for

- Power interactive phone and voice bots with sub-second response times
- Clone a custom brand voice using a short 10-second reference clip
- Localize audio content and voice tracks across 44 languages while maintaining identity
- Insert realistic non-verbal cues like laughter or breaks directly into synthesized transcripts

## Features

- Sub-90ms first-audio model latency for streaming audio playback
- Instant voice cloning from short audio samples (5 to 10 seconds)
- Support for 44 languages and diverse regional accents
- Inline emotion, pacing, and non-verbal expression control (such as laughter markers)
- Custom pronunciation dictionary support for domain terms and acronyms
- Enterprise deployment options including Cloud API, VPC, and on-premises/air-gapped setups

## Pricing

Freemium. Free tier with 20K credits; paid self-serve and custom enterprise tiers available.

## Details

- Type: AI model
- Category: [AI](https://xiry.app/topics/ai.md)
- Topic: [Text to Speech](https://xiry.app/topics/ai/text-to-speech.md)
- Open source: no
- Platforms: Web, API, Self-hosted
- For: Developers and teams building real-time voice agents, customer support bots, and interactive voice apps.
- Alternative to: ElevenLabs, PlayHT, Deepgram, OpenAI TTS
- Tags: tts, voice cloning, speech synthesis, voice ai, audio, api

## Why it was saved

- [@svpino, 8 February 2025](https://x.com/svpino/status/1888320663200604319): Free to try; hallucination-free, uncannily similar to the real voice, high speed and very low latency — practical for voice agents and products with voice.
- [@LinusEkenstam, 29 October 2025](https://x.com/LinusEkenstam/status/1983419340591640966): SOTA voice cloning from a 9-second clip, sub-second latency, inline audio instructions (\<speed ratio\>, \<break time\>), laughter insertion, emotion and speed controls — practical reference for realistic voice AI agents.

## Compare

- [Alternatives to Play.ht](https://xiry.app/alternatives/play-ht.md)
- [Alternatives to ElevenLabs](https://xiry.app/alternatives/elevenlabs.md)

## Related

- [ElevenLabs](https://xiry.app/resources/elevenlabs): AI voice generator, voice cloning, and conversational agent platform (Web app, Freemium)
- [Dia](https://xiry.app/resources/dia): Open-source 1.6B text-to-speech model generating multi-speaker dialogue in one pass. (AI model, Open source)
- [Airy](https://xiry.app/resources/airy): Fast, low-cost text-to-speech API and studio for creators and developers (AI model, Freemium)
- [AudioPen](https://xiry.app/resources/audiopen): Converts spoken thoughts into structured, polished writing in customizable styles. (Web app, Freemium)
- [cartesia-python](https://xiry.app/resources/cartesia-python): Official Python library for Cartesia's real-time text-to-speech API (GitHub repo, Paid)
- [Chatterbox](https://xiry.app/resources/chatterbox): Open-source zero-shot text-to-speech and voice cloning model (AI model, Open source)

Updated 10 October 2026. Page: https://xiry.app/resources/cartesia-sonic

Make your own library: Xiry takes your saved reels, tweets, Reddit bookmarks and Watch Later videos, makes them searchable and pulls the useful stuff out, like this library. It's invite-only for now; join the waitlist at https://xiry.app/#waitlist.