Piper TTS

Designed to run entirely on local hardware, Piper is a fast neural text-to-speech engine that works even on single-board computers like the Raspberry Pi 4. It pairs compact VITS-based voice models with espeak-ng for phonemization and offers many voices across dozens of languages, plus documented support for training custom voices. Originally developed within the Rhasspy voice assistant ecosystem, the project now lives under the Open Home Foundation as piper1-gpl and carries a GPL license. It can be used as a command-line tool, an HTTP server, a Python API, or a C and C++ library, and operates fully offline with no cloud dependency, which matters for privacy-sensitive and embedded deployments. Adoption is broad: Home Assistant uses it for local voice, the NVDA screen reader uses it for accessibility, and it appears in the Open Voice Operating System along with education and gaming projects. Users are typically home automation enthusiasts and developers who need offline speech output.