Good day!
PDFs work stably, please, tell us about the issues you are having with them. The case we did not support but will roll out this week is OCR support for scans.
YouTube worked perfectly fine before they introduced the authorization wall a couple of weeks before our launch because Nvidia was scraping 80 years of videos per day. We're still grabbing transcripts, but might be not as smooth as before when we downloaded audio and transcribed it with Whisper.
We regret that you encountered issues, you can always make a refund, or wait a bit for us to fix things that did not work – we're a startup and running as fast as it is possible.