Moshi by Kyutai

Moshi by KyutaiSince 2024-07

Moshi is an open-source real-time conversational voice AI foundation model developed by Kyutai, the non-profit AI research lab based in Paris. Engineered to revolutionize human-AI verbal communication, Moshi operates on a full-duplex architecture capable of listening, thinking, and speaking simultaneously with sub-200ms end-to-end latency. Unlike traditional voice assistants that chain separate Speech-to-Text (STT), Large Language Model (LLM), and Text-to-Speech (TTS) pipelines together, Moshi processes raw multi-stream audio natively as continuous speech tokens. This allows Moshi to understand emotional nuances, interrupt and be interrupted naturally, chuckle, whisper, and express genuine conversational timing. Moshi is fully open-source with openly accessible weights, training recipes, and inference code, serving as a foundational milestone for research in real-time spoken language modeling and multi-modal conversational systems.

Moshi’s architecture is built on Helium, a 7-billion parameter language model coupled with Mimi, a cutting-edge neural audio codec that compresses 24kHz audio into multi-stream discrete tokens at just 1.1 kbps. By operating on a joint text-audio token stream, Moshi predicts both conversational text tokens and acoustic speech tokens in parallel. This end-to-end audio modeling eliminates the latency bottlenecks and acoustic information loss inherent in cascading STT-LLM-TTS pipelines. The Kyutai team provides full PyTorch and Rust-based inference engines optimized for local GPU execution, enabling real-time full-duplex conversations on consumer-grade hardware (NVIDIA RTX 4090 or Apple Silicon Mac).

CartesiaSince 2024-04

Cartesia is an AI audio and voice intelligence company that powers conversational AI agents and interactive applications with ultra-low-latency, hyper-realistic voice synthesis. Its flagship State Space Model (SSM) architecture, Sonic, delivers human-like voice generation with sub-90ms time-to-first-audio latency. While traditional transformer-based text-to-speech models struggle with high latency and compute overhead, Cartesia's lightweight SSM architecture allows developers to build fluid, conversational voice bots that feel instantaneous and natural, without awkward pauses. Cartesia is used by developers, contact centers, game studios, and AI agent builders across the world to power real-time phone assistants, gaming NPCs, live transcription translation, and voice-enabled enterprise interfaces.

Cartesia's Sonic engine is built on state space foundation models designed specifically for continuous streaming audio. It supports multilingual voice cloning from a 5-second sample, fine-grained emotional control (whisper, excitement, professional demeanor), and dynamic pacing adjustments. Developers can integrate Cartesia via WebSocket and REST APIs, Python and TypeScript SDKs, and WebRTC streaming for mobile and browser applications. The platform handles concurrent high-throughput workloads with strict SLA guarantees.