OpenAI Whisper endpoint doesn't transcribe the audio properly

API Bugs
1

eg:
What I asked "hi can you assist me "
What it assumed “Hi Kenyas speak to me”

Sometimes, getting random words which is not part of my speech