---
id: "b901"
language: "en"
library: "tools"
official_url: "https://github.com/kyutai-labs/moshi"
profile_updated_on: "2026-10-20"
---

# Moshi

Model user and assistant audio streams together for full-duplex spoken dialogue, with PyTorch, MLX and Rust inference implementations.

- Company / organization: Kyutai
- Company country / region: France · Company headquarters
- Category: Audio & music
- Official website: [https://github.com/kyutai-labs/moshi](https://github.com/kyutai-labs/moshi)
- Profile updated: 2026-10-20


## Overview

Moshi jointly models user audio, system audio and text, allowing listening and speaking to overlap instead of waiting for a complete recording to be transcribed. It uses the Mimi streaming codec, with PyTorch research code, Apple-device MLX, a Rust backend and web client. Hardware and quantization vary by implementation. Python and client code use MIT, Rust uses Apache, and official weights use CC-BY 4.0.

## Official sources and references

- [Official website](https://github.com/kyutai-labs/moshi)

## Reading & subscriptions

- [Interactive profile](../../../products/moshi--b901.html)
- [Read](../../../read/en/products/moshi--b901.html)
- [JSON](../../../read/en/products/moshi--b901.json)

Descriptions summarize public information; consult official websites for features, pricing and regional availability.