Phonon
Open speech recognition models for English, running on a laptop or datacenter GPU.
Servicefreeglobal

Phonon is a set of free, open speech recognition models for English from Fermion Research, small enough to run on a laptop or a datacenter GPU with a CLI and CPU/CUDA images, transcribing an hour of audio in about two and a half minutes. It's aimed at developers who want local, open speech-to-text, similar in role to Whisper, instead of relying on a cloud speech API.
Categories
ai-infrastructuredeveloper-tools
Something wrong with this listing — dead link, not a real product, wrong info?

