I find that the transcription is quite accurate when I manage to get some transcriptions. I can also get a nice .srt file for my close captions.

However, there are two issues that makes this service a bit inconvenient:

1- It is often down. Luckily, I don't often have tight deadlines so I can wait a few days, but it is inconvenient and might cause issues for someone who works with short deadlines.

2- When I get the full text, I wish it would split that into some logical paragraphs. Since I don't need the [speaker] (I am the only one talking in the video) or the [timestamp], I end up with a big block of text that could be 10 pages long. It might not sound like much but it requires me to read every word in order to split the text into legible paragraphs for my users. I am sure, with AI, there is a way to do that so the document is more reader-friendly.