Why turn a document into Markdown at all?
Markdown keeps what a document says and how it is organised: headings, paragraphs, lists, tables, links, emphasis. It drops how the document looks: fonts, margins, colours of the letterhead, the exact place a logo sat. For a lot of work that is the right trade.
Searching. A folder of .md files is searchable with any tool, including the one built into your operating system. A folder of PDFs and scans is not.
Reading on anything. Plain text renders on a phone, in a terminal, in an email, in twenty years’ time.
Feeding it to something else. Notes apps, static-site generators, wikis and AI assistants all read Markdown and choke on .docx. A contract turned into Markdown can be quoted, diffed and summarised.
Editing without a licence for the original program. You do not need Excel to fix a number in a converted spreadsheet.
Version control. Markdown diffs line by line. Two versions of a Word file do not.
The trade only works if the conversion is faithful, and that is where most tools give up. Pasting from a PDF viewer gives you one line per visual line with hyphens in the middle of words. “Export to text” loses the tables. Anymark’s PDF engine rebuilds reading order across columns, rejoins paragraphs, finds the headings by size and position (never mistaking small print for the body), and keeps tables as tables whether they have rules or not. Scans get the same treatment after OCR.
Where it does not work: anything whose meaning is in the layout. A poster. A complex form with arrows. Mathematical typesetting. Anymark will give you the words, but you will want the original beside you. We say so in the app rather than pretending.