Phonon is a 65 to 75 MB on-device text-to-speech model that runs locally on phones laptops and browsers without network access. It uses flow matching architecture to generate audio in English French German Spanish and Portuguese with over 100 built-in voices.
The 31 million parameter version is Apache 2.0 licensed making it suitable for commercial products while larger versions are restricted to non-commercial use. It tops CPU streaming benchmarks for speed but you must handle voice cloning integration separately if you need features beyond the built-in voices.
more: https://huggingface.co/blog/phonon-tts
my other side: ivo.blogspot.com
Leave a Reply