Overview
Piper is a local neural text-to-speech engine designed for fast inference on personal computers and embedded hardware. The project provides voice models at several quality levels, including single-speaker and multi-speaker options, and supports generating audio without sending text to a hosted service. It can be used through command-line tools and software integrations, with voice samples and model information published for selection. The active source repository is maintained by OHF-Voice and is licensed under GPL-3.0. The earlier rhasspy/piper repository is archived, so the current project is marked moved.