Open CS1o opened 1 year ago
I just had a look at the vosk libraries used by Sayboard. It's about 8 megabytes for each CPU architecture, so this alone would add a considerable chunk to app size.
What is the benefit of implementing vosk STT directly in the app, over having the voice input button switch to Sayboard? Only that everything is within a single app, or is there anything more?
What is the benefit of implementing vosk STT directly in the app, over having the voice input button switch to Sayboard? Only that everything is within a single app, or is there anything more?
Describe the solution you'd like Adding Vosk Speech to Text (STT) Support for the Voice Input Button for Devices with no Google Services installed.
Use case In using a Pixel 7, Android 13 with GrapheneOS installed. (No google services installed). After the v10 update we now have the Voice Input Icon. (This icon is not shown for GrapheneOS. I think its only shown if a STT backend like Vosk or Gboard voice model is found). I would like to use the STT for Apps like Signal. Therefore Openboard would need its own Vosk Integration/support.
Describe alternatives you've considered I've installed the foss app Sayboard from github. In the App i granted the permissions and downloaded the Vosk language file. After that the Voice Input Icon appeared in Openboaed and TTS works.