tools
Moshi
Moshi is an open-source speech-to-speech model for real-time, full-duplex voice conversations.
In inglese
Moshi listens and speaks directly through audio streams instead of using separate speech recognition, text generation, and text-to-speech stages. It models the user’s and assistant’s audio in parallel and produces an inner-monologue text stream to guide responses.
Researchers and developers can run it locally through PyTorch, MLX, or Rust backends, use its web interface, or adapt the code for experiments. It speaks English and requires substantial hardware for some configurations; Windows is not officially supported.
Features
- Full-duplex spoken dialogue with simultaneous listening and speaking
- Direct speech-to-speech generation with parallel audio streams
- Inner-monologue text stream to improve response generation
- Streaming Mimi neural audio codec for 24 kHz audio
- PyTorch, MLX, and Rust/Candle inference backends
- Local web interface and command-line client
- Three released model variants, including Moshika and Moshiko
- Code under MIT and Apache-2.0 licenses; model weights under CC-BY 4.0
Use cases
- Build low-latency voice conversation prototypes
- Research speech-native language models
- Run voice dialogue locally on Mac or Linux hardware
- Fine-tune voice dialogue behavior for experiments
- Create custom web or command-line speech interfaces
Pros
Cons
Latest updates
- mlx-0.3.0: Add batching support to the MLX implementation (#347) (mlx-0.3.0)
The MLX implementation adds batching support for STT.
- mlx-0.2.11: Moshi mlx.0.2.11 (#333) (mlx-0.2.11)
The MLX local-web example was fixed, ported to sphn 0.2, and the Moshi MLX dependency was bumped.
Pricing
- Starting price
- Free
- Prices checked
- 2026-09-25
- Read
- by web search
Free / Demo
Free
- Real-time full-duplex voice conversation (sub-200ms latency)
- Experimental conversational AI prototype (non-commercial use)
- Limited to 5-minute conversations
- Open-source models (Moshi, Mimi codec) available for self-hosting
- Access via online demo at moshi.chat / kyutai.org