firecrawl/pdf-inspector ? reverse-engineered prompt

Build me a fast PDF inspection tool in Rust that can tell whether a PDF is text based, scanned, image based, or mixed, then extract the text in the right reading order and turn it into clean Markdown.

I want it to work well on reports, papers, invoices, and legal PDFs, especially ones with tables, headings, lists, and multi column layouts. It should also give a confidence score for the PDF type and basic page by page routing so callers can decide when OCR is needed.

Please include a simple command line tool that can convert a PDF to Markdown, output JSON when asked, and do detection only if needed. It would be great if the core library could be used from Rust, Python, Node, and the browser too, but keep everything local and fast, without relying on external services. If you need to check current docs or examples online, go ahead.