I'm pretty much exclusively a Linux user and when I did my OSCP (24 hour infosec exam), you were not allowed to screen record the exam, but you were allowed to take excessive screenshots (such as 1 per second) .<p>Screenshots were essential for the post-exams report / write up that you had to submit the following day.<p>Rather than have to rely on taking screenshots, and my bash history I did exactly this, including OCR to help me search for the relevant image. It was super useful, and for this sort of requirement a perfect fit.<p>Recall does have a usefulness, but it's just been implemented poorly so far.
Most knew you can do screenshots automatically, small number of people knew easy text recognition is widely available (I am not one for instance), but this lacks text indexing and AI-based querying. So this is as much a clone of Recall as a bicycle would be of Toyota RAV4.