[frontrowseat]

> Notice: This is a demo server. Text to speech is expensive. You will be able to add new articles until I have spent $300. This server will be up for about a week.

If this is offline first, then why use a server at all? Why not use the native OS text-to-speech? MacOS has built in text-to-speech and so does Windows (Narrarator?)

[doomrobo]

Good question. The short answer is that the UX of doing it with local TTS is pretty terrible. The voices will of course vary from device to device. But more importantly, control from the lockscreen doesn't work anymore (I've found you can't really use the web TTS API in a way that plays well with the MediaSession API, the position-resuming functionality, etc)

[selfhoster11]

Browser TTS on Linux is really terrible, for one

[mbreese]

What about Android?

We're really talking about iOS and Android here, if I grok the use-case.

[doomrobo]

Try it out here https://mdn.github.io/dom-examples/web-speech-api/speak-easy...

It's god awful on iOS

[obviyus]

Some of those are nightmare inducing.

[selfhoster11]

Android is not really Linux here. It normally includes the proprietary TTS engine from Google, which is quite decent. Even offline, it’s many times better than what Firefox on Linux comes with.