Moshi

by Kyutai

Kyutai's open speech-to-speech foundation model and demo voice agent.

TL;DR

Kyutai's open speech-to-speech foundation model and demo voice agent.

Best for researchers experimenting with full-duplex speech-to-speech agents. Pricing: free.

Category
Open source
License
—
Stars
—
Last push
—
Pricing
free
Platforms
Linux, macOS, Cloud

What it is

Moshi from Kyutai Labs is an open-weights full-duplex speech-to-speech foundation model: it listens, thinks, and speaks simultaneously rather than turn-taking. The released checkpoints and demo agent are useful for exploring how non-cascaded speech models change voice-AI latency budgets. Apache-2.0 licensed.

Best for: Researchers experimenting with full-duplex speech-to-speech agents.
Watch out for: Research grade; English-led; not a hosted production API.

Features

Speaker diarizationNo
Word-level timestampsNo
Streaming / real-timeNo
Languages supportedNone
HIPAA eligibleNo

Moshi vs Whipscribe

FeatureMoshiWhipscribe
CategoryOpen sourceTranscription APIs
Pricingfreefree beta
Speaker diarization—Yes
Word timestamps—Yes
Streaming—No
Languages—99
PlatformsLinux, macOS, CloudWeb, API, MCP

Alternatives to Moshi

OpenAI Whisper
OpenAI

The reference open-source multilingual ASR model from OpenAI.

OSS · MIT ★ 98.1k
whisper.cpp
Georgi Gerganov

C/C++ port of Whisper — runs on anything, from a Raspberry Pi to Apple Silicon.

OSS · MIT ★ 48.8k
faster-whisper
SYSTRAN

4× faster than reference Whisper using CTranslate2 — production sweet spot.

OSS · MIT ★ 22.3k

Whipscribe is a managed faster-whisper + whisperX service. If you want transcripts without running infrastructure, paste a URL or drop a file in the form below — you'll have a transcript in seconds.