It uses the VOSK API to record audio on the device, transcribe it, and display it as subtitles.
It can be used to transcribe instead when sound is not available.
When you start the app for the first time or when you add other languages, you will need to download the model files required for transcription.
https://alphacephei.com/vosk/models
It uses screen recording to record. It only accesses the audio.
It also uses a foreground service to keep the transcription running in the background.
This app is open source
https://github.com/takusan23/Hiroid