Been losing very valuable content when I talk to openAI whisper and the audio exceeds character limit (usually only by like 2-3000 tokens!!) it would be super helpful to have some error handling or what have you.