Hugging Face introduced Phonon, a new text-to-speech model designed to run entirely on a device without an internet connection. The release is open source under the Apache 2.0 license, allowing developers to use the technology freely. Phonon includes a new inference framework called ptts that supports various hardware backends.
The model comes in two sizes, approximately 65 MB and 75 MB, and supports five languages including English, French, German, Spanish, and Portuguese. It offers more than 100 built-in voices and currently ranks first on Hugging Face’s Open TTS Leaderboard for German and Spanish. The project has been downloaded over 200,000 times in its first two weeks.
This tool lets creators build voice applications that work offline, which is useful for privacy-sensitive or low-bandwidth environments. The small file size and hardware support make it practical for mobile and embedded devices. You can integrate high-quality speech generation into your own projects without relying on cloud services.
