[ STACK LIBRARY ]

Kyutai

Kyutai

Visit website

Speech-to-Speech

ABOUT

Paris nonprofit AI lab behind Moshi, the open-source speech-to-speech model that talks and listens at once.

Kyutai researches open science AI with a focus on expressive real-time conversational voice. Its flagship model Moshi listens and speaks simultaneously through a 7-billion parameter multimodal language model paired with the Mimi neural audio codec, reaching latency around 160 to 200 milliseconds. Kyutai releases Moshi's code and weights as open source, and also built Hibiki for simultaneous speech-to-speech translation.

STACK COVERAGE

Speech-to-Speech

Open-source models & research, Nonprofit AI lab

DETAILS

HQ COUNTRY

France

FOUNDED

2023

LANGUAGES

English, French

DEPLOYMENT

Open source, Self-hosted

PRICING MODEL

Open source (Free)

COMPANY STAGE

Nonprofit

Is this your company? To claim this listing or update its details, contact us at sunil@uniocommunity.com

UNIOVOICE AI STACK LIBRARY
Kyutai logo
Trusted by Unio

Kyutai

Speech-to-Speech

Paris nonprofit AI lab behind Moshi, the open-source speech-to-speech model that talks and listens at once.

Open-sourcemodels&research,NonprofitAIlab
Layer
Speech-to-Speech
Market
Global
HQ
France
Founded
2023
Deployment
Open source, Self-hosted
Pricing
Open source (Free)
Languages
English, French
Stage
Nonprofit
Listed on the Unio Voice AI Stack Library

More in this layer

var(--variable-H7lUBNgYB)

.wave

Speech-to-Speech

Continuous inference infrastructure for real-time voice and AI models, backed by Y Combinator.

var(--variable-H7lUBNgYB)

Achha AI

Speech-to-Speech

Multilingual text, speech, and image processing through a multimodal AI platform.

[

India-ready

]

var(--variable-H7lUBNgYB)

Amazon Nova Sonic

Speech-to-Speech

Speech-to-speech foundation model on Amazon Bedrock for building real-time conversational voice agents.

Foundation model on Amazon Bedrock

[

India-ready

]