Carsonthemonkey
commented
1 year ago The problem was actually with the debug text. It had excessive newline characters that would not be generated by the actual transcript. I have replaced it with one that is more simple, but if we add the ability to upload text files, or use a more advanced transcription method we may need to take another look at this issue.
Giving over 700 when about 300 words shown on screen (however, this is in debug mode so it might work with the API calls)