First usable prerelease

This pre-release lacks testing on other devices, at the time of writing the app was tested only on my Samsung Galaxy M21 running android 12.0 (API level 31), elsewhere may crash. The cache management is currently absent and the UI is pretty spartan, but you still get the ability of being able to share your Whatsapp or Telegram message (or any audio file actually) to VolMan so it gets transcribed, the model is also selectable since the way of managing the model has been heavily improved architecture wise, also right now the models come with the app itself, mandatory online download has been removed (so we get better privacy as well) and the user can choose between the 2 models currently available.
one last feature that has been added is the ability of opening the Github repo where licenses and code can be seen (this way we comply with the various library licenses).

Rough Demo!

This application is far from stable and needs to be optimized UI wise (I put everything in the main thread but I was actually just testing), in the end it just works, so if you want to play around with it here we go.
at the moment the default model is whisper-base which is not very fast but still works well enough and doesn't make many mistakes.
remember to press the download model button that will download the model from huggingface
note: the application always encodes the file via libvlcsharp into wav, libvlc should support about anything, if it crashes because of that let me know.