kyutai

A Docker-based OpenAI-compatible Text-to-Speech API server powered by Kyutai's TTS models with GPU acceleration support.

Private, on-device meeting transcription and dictation for macOS. Local STT, summaries and action items, full exports, MCP server for AI assistants. No cloud, no API keys.

Kotai is a fully local, zero-cost voice assistant that combines the power of Kyutai TTS/STT, LiveKit, and local LLMs to create natural conversational experiences.

Audio to MIDI ComfyUI Node

An automated installation script for deploying Kyutai's Moshi STT server on macOS Apple Silicon.

A FastAPI-based Speech-to-Text service that provides OpenAI Whisper API compatibility using Kyutai's powerful STT models. This allows you to use any OpenAI Whisper client with Kyutai's models as a drop-in replacement.

Demo repository for Kyutai Labs' STT-1B model: Real-time speech-to-text transcription with streaming inference, built-in VAD, and Jupyter notebook examples for audio processing and simulation.

LiveKit TTS plugin with Kyutai streaming implementation

Golang bindings to Kyutai Delayed Streams Modeling Rust productions servers

A high-performance, GPU-optimized real-time speech-to-text (STT) streaming server built with WebSocket support for multiple concurrent clients. This project leverages the Kyutai STT model and is optimized for NVIDIA RTX 4090 GPUs, providing low-latency transcription for audio streams.

Agent skill that speaks replies aloud with Kyutai Pocket TTS - fast time-to-first-word, runs on CPU with zero config, clone any voice

Free, offline text-to-speech for conversations — give it a transcript, pick a voice per speaker, get one audio file. Runs on CPU.

FastMCP 3.1 server plus webapp for kyutai moshi speech tool

A FastRTC wrapper for the Kyutai Pocket TTS model

VoxReach POC — full-duplex AI receptionist demo built on Kyutai MoshiRAG (open weights). Persona: Hearth & Pass Korean restaurant. 2-pane investor demo UI + POS-write moat story.

Local voice assistant built with LiveKit, Kyutai STT/TTS, and Ollama

Working integration with Kyutai and the Omi app.

Local test bench for Kyutai Pocket TTS — a 100M-param, CPU-only, MIT text-to-speech model. Pick a voice and hear it synthesized on your CPU, or start a live voice call. Next.js UI + Python sidecar.