When using talkscriber, it does a fantastic job of multilingual transcription, but a majority of the transcripts appear to be in english. From time to time it will be in the language spoken, but the transcript itself is often translated to english. Is there a way to keep the transcripts in the original spoken language? This way the ai model used is able to determine the language spoken and give responses in said language. Basically allowing it to adapt to the spoken language. Speaking to an ai in japanese and having it respond in english isn't terribly useful, despite the transcription model understanding everything