i have feature request for streaming translations. i attend meetings where everybody speaks in language i do not understand. started using superwhipser to translate. currently i have to record small piece, end recording and then processing takes a bit time before i see translation but i cant record new piece while processing takes place. so if there were a feature with sort of streaming/buffering translations that would be great. additionally if it would read out the streamed translationt that would be even more amazing. before i found out about superwhisper i started to make such program myself. spend few hours messing with cursor-ai and openai whisper api. got it somewhat working but it was really poor detecting when talking starts so i gave up trying to do it myself.