Hi Paul,

What exact issue do you face? What are your OS and Firefox versions?

Firefox, for some reason, doesn't support speech recognition. It is a javascript interface that Chrome and Edge fully support. I don't think it matters what version. I happen to have Windows 10 and 11 systems. My Firefox must be the latest version as it was updated recently.

The reason is existing implementations only rely on uploading the audio to a cloud service for transcription. (Read: to Google…)

The more interesting workstream is for processing locally, on–device, but that's still in early stages…

I am talking about someone talking in order to update a textarea. There is no need for an upload to anything.

Google chrome works perfectly every time. Edge has an occasional hoccup with a "network" error that I have no answer for.

I am talking about someone talking in order to update a textarea.

Yes. SpeechRecognition API. See https://webwewant.fyi/wants/55/

The WebSpeech draft https://webaudio.github.io/web-speech-api/ is where the on–device debate is happening: https://github.com/w3ctag/design-reviews/issues/1038

There is no need for an upload to anything. […] Google chrome works perfectly every time. Edge has an occasional hoccup with a "network" error that I have no answer for.

Eeer… because, Chrome uploads the recording to Google. And Edge most likely too.

Here's the standards-position+ on the latest proposals and further links for following any progress:

• https://mozilla.github.io/standards-positions/#web-speech-api-on-device