Kyutai Labs Releases Pocket TTS, a CPU-Efficient Text-to-Speech Application
Kyutai Labs has launched Pocket TTS, a lightweight text-to-speech (TTS) application designed to run efficiently on CPUs, eliminating the need for GPUs or web APIs for TTS models. This application supports Python versions 3.10 through 3.14 and requires PyTorch 2.5+. Key features include a small model size (100M parameters), audio streaming, low latency (approximately 200ms for the first audio chunk), and faster-than-real-time performance (around 6x real-time on a MacBook Air M4 CPU). Pocket TTS uses only two CPU cores, offers a Python API and CLI, and supports voice cloning and multiple languages including English, French, German, Portuguese, Italian, and Spanish. It can also handle infinitely long text inputs and run client-side in a browser.