PDF to Markdown

No. This is the one conversion in the family that never calls pandoc: pdf.js extracts the text in a Web Worker and the Markdown it produces is the finished file. There is no upload, no server step and no account, and the whole run works offline.

By glyph height, not by any structure in the file. It counts how often each text size appears across the document, calls the most common one body text, and treats every size more than 15 percent taller as a heading, ranked so the largest becomes h1. A line starting with a dash, asterisk or bullet becomes a list item. It is a heuristic, so a document with no size contrast comes out flat.

To keep your document's own text intact. A PDF has no Markdown semantics, so every character Markdown gives meaning to is escaped: backslash, backtick, asterisk, underscore, brackets, angle brackets, ampersand, dollar, pipe, tilde, caret and braces, plus block markers at the start of a line. Without this, a path like C:\temp\notes.txt was read as TeX and came out as C:.txt.

By vertical position. pdf.js reports each run of same-font text separately, so items whose y coordinates are within 2 points of each other are grouped into one physical line, sorted left to right, and the lines are ordered from the top of the page down. Pages are concatenated with no page break marker between them.

A .md named after your PDF, with the Markdown also shown on the page as a preview. The preview is byte-identical to the downloaded file, so what you read is what you save. One PDF at a time, and the original is untouched. If there was no text to extract, it stops with "no extractable text found in this PDF".

PDF to text gives you the raw content with a marker before each page and no interpretation. This one groups items into lines, guesses headings and lists, and escapes the result so it renders back as the words on the page. Pick this when you want something readable, and PDF to text when you want something to search.