Readlink

The "ask it like you'd ask a friend" search is a great angle. A lot of what I need to find later is a passage I read inside a PDF or ebook, not just the file itself. Does Trailback index the text of the pages you actually read in a PDF, so "that paragraph about habit loops from last week" would work, or does it match on the file and its title? And since it captures screenshots, can you exclude specific apps or sites from being recorded?

The PDF to Markdown parsing is the part I'd use most. We deal with a lot of book PDFs, and the hard cases are scanned pages and two-column layouts where footnotes end up mixed into the body text. How does the parser handle reading order on multi-column pages, and does it run OCR on scanned PDFs or only work with ones that already have a text layer?