How does it work on a technical level?

There’s virtually no information on the experiment page.

Considering that this is quite sensitive topic, it would be nice to have a broad overview of how this works.

How and where do you process the voice? Locally? On the server?

If later then whose services do you use? If former, then how voice is being processed to speech? Do you bind to the OS TTS services? Do you use 3rd party services. Did you build your own?

What I want to know is where my full voice data goes, and who has access to the text besides me and my local machine.


I’d like to add my +1 to this data privacy / security question.

I’d love to use this extension but until I know that my voice data is not is going to third-party servers who might sell my data (e.g. Google et al), I’m not going to take the risk.

Is everything processed locally in my browser / OS? If so, great. If it isn’t then whose servers is it processed on and what are the privacy and security protections in place?

From the test pilot page related to voice fill: Here

When you choose to use Voice Fill, your recordings will be sent to our servers for processing to provide you with suggested text.

There was a further link with some detail on the data collection. I’d like to try it but I’d only really like to use voice on mobile and there experiments don’t seem to work for Firefox beta on mobile.