Markdown to PDF

What renders

The converter reads GitHub-flavoured Markdown, the dialect used by GitHub, most docs sites and most note apps, and sets it as real text, so the PDF can be searched, copied and read by a screen reader.

• Six heading levels, each set larger and bold, so the outline of a README or a spec is visible at a glance.
• Bold, italic, strikethrough and inline code, with inline code set in a monospace face on a light grey background.
• Fenced and indented code blocks keep every space and line break, in monospace on a shaded panel. Long lines wrap rather than running off the page.
• Tables become ruled grids with bold header rows, sized to the page width.
• Bulleted, numbered and nested lists, including numbered lists that start from a number other than one.
• Task lists render with real checkboxes, ticked or empty, rather than the raw [x] and [ ] characters.
• Block quotes are indented with a bar down the left edge, and nested quotes get one bar per level.
• Horizontal rules, and web and email links that stay clickable in the PDF.

What is deliberately left out

• Images that point at a web address or a file path are not downloaded. Fetching one would tell that server the document was being opened, and a file path means nothing to a browser. The alt text is printed in its place and the tool tells you how many it skipped. Images embedded directly in the file as data are included.
• Raw HTML that can run code or load something, such as scripts, styles, iframes and forms, is removed before anything is laid out. The text inside ordinary HTML tags is kept.
• A YAML front-matter block at the top of the file, the --- metadata section used by Jekyll, Hugo and Obsidian, is removed, because it is settings for the publishing tool rather than part of the document.
• Syntax highlighting is not applied to code blocks. They are set in a single colour so they print well in black and white. For colour-highlighted source code, the Code to PDF tool is the better fit.
• Maths written in LaTeX, Mermaid diagrams and footnotes are app-specific extensions rather than core Markdown, so they come through as their source text.
• Links to other Markdown files or to headings in the same file are not clickable, because there is nothing for them to point at once the document is a single PDF.

Getting a good result

• Body text is set in Helvetica and code in Courier, two of the fonts built into every PDF reader, so nothing is downloaded and the file stays small. Both cover English and Western European languages, including accents, curly quotes, dashes and the euro sign.
• Characters outside those fonts, such as Greek, Cyrillic, Polish letters, arrows, box-drawing lines in a directory tree, and emoji, are replaced with a question mark. The tool tells you how many it replaced, so nothing changes silently.
• Pick A4 or US Letter to match where the PDF is going. Wide tables are squeezed to fit the page, so a table with many columns reads better in fewer, wider columns.
• Pasting is handy for a quick note or an answer copied from a chatbot. The file name of the PDF comes from the first top-level heading.
• Skim the PDF once before sending it, paying attention to wide tables and very long code lines, which are the two places the page width shows most.
• Everything happens in your browser. Internal notes, unreleased documentation and anything under an NDA never leave your machine.