Speech-to-Text Client-Side?

[novantadue]

Is there any way to do conversational speech-to-text in the browser/with mobile GPUs? Or do we need to still choose a cloud provider?

[zcw100]

Give Vosk-Browser a try https://github.com/ccoreilly/vosk-browser

You can also try kaldi-wasm but it hasn't been worked on for a while https://gitlab.inria.fr/kaldi.web/kaldi-wasm

[Dasky]

I think speechly is working on that: https://www.ycombinator.com/companies/speechly