Tagged PDF, explained
An institution asked for a tagged or accessible PDF. This page is the short version: what they actually check, why files fail, and the smallest fix in your Markdown.
MarkdownGo starts from Markdown: the free check reads pasted text, a .md file, or an exported PDF; the new PDF is built at export. It does not repair an existing PDF — a scan or a file made in another tool needs OCR and manual tagging.
What a tagged PDF actually is
A PDF is a pile of drawn glyphs. Tagging adds a structure tree on top of it: this block is a heading, that run of text is a paragraph, this image is a figure with alternative text. Screen readers follow the tree, not the pictures of text. Without tags, a reader has to guess and usually reads the whole document as one long blob.
PDF/UA is the standard that defines a tagged, accessible PDF. PDF/A is about long-term archiving — it is not the same thing, and having one does not give you the other.
What checkers look for
- A document title — the name assistive tech announces, not the file name.
- A document language — so the reader knows whether to speak English or Chinese.
- A heading hierarchy without skips — h1, then h2, then h3. Never h1 straight to h3.
- Alternative text on every figure that carries meaning; empty alt on purely decorative images.
- A reading order that matches the source, plus bookmarks for long documents.
Fixing each one in Markdown
- Title: set a real title in Front Matter (`title:`), or keep exactly one H1 at the top of the file.
- Language: put `lang: zh-CN` in Front Matter, or let the exporter detect it from the text. The language of the interface you are reading does not decide it.
- Headings: one `#`, then `##`, then `###`. Skipping a level is the single most common failure.
- Alt text: `` for charts; `` for decoration.
- Reading order: write in the order you want it read. Side-by-side HTML blocks are the usual culprit when it is wrong.
Verify before you submit
Run the free check first: it reads your Markdown or an exported PDF and reports the source-level failures above — title, heading skips and alt text — plus which ones the exporter fixes for you. The document language comes from `lang:` in Front Matter, or from automatic detection — not from the interface language. Reading order and bookmarks are properties of the exported PDF, so check them in your reader after export.
Short answers
Does PDF/A make a file accessible?
No. PDF/A governs archival durability (fonts embedded, no external dependencies). Accessibility comes from tags, which is PDF/UA.
Do I need Acrobat to check a PDF?
No. The free check works from your Markdown before you export, and you can also upload an exported PDF to check it; any PDF reader that lists the structure tree (or a free checker) can inspect the result afterwards.
Will my university accept the export?
Ask which standard they require (PDF/UA, WCAG-tagged, or their own checklist). If it is PDF/UA, turn on the PDF/UA switch — but the export still has to pass the checks above and your institution's own review.
Free PDF/UA check. PDF/UA and PDF/A exports are part of Pro.