Images are missing alt text
Missing alt text on some images (3 in total). A PDF/UA export is refused until every image has a description.
Free accessibility check
Paste a Markdown document or upload an exported PDF, and run the same PDF/UA checks the exporter runs: missing alt text, skipped heading levels, empty links and a missing document title. Nothing is stored.
Free · No sign-up · Nothing stored
Checked in memory. Your document is not stored — a share link keeps only the issue summary, and only if you create one.
What the check covers
Four source-level issues stop a Markdown export from passing PDF/UA: an image with no alt text, a heading level that skips (H1 straight to H3), a link with no target, and a document with no real title. The check points at the line in your Markdown; only errors block the export.
Missing alt text on some images (3 in total). A PDF/UA export is refused until every image has a description.
“Results” jumps from H1 to H3; PDF/UA export normalises it to adjacent levels.
Link “the appendix” has no target address.
PDF/UA export falls back to a fixed title; add a Front Matter title or an H1.
How it works
Drop in a .md file or paste the text — nothing leaves your browser until you press check.
The same rules the exporter uses: images without alt text, heading levels that jump, links without a target, a document without a title.
Heading levels and the title are normalised at export time; alt text is a one-line fix in your Markdown.
Questions
No. Checking, previewing and re-checking are free; only exporting a finished PDF counts against the daily export limit.
Yes. Switch to the Exported PDF tab and upload the file — the check reads its structure tree, embedded fonts, title, language and alt text.
Missing alt text, skipped heading levels, empty links and a missing document title — the same rules the PDF/UA exporter uses. The check points at the line in your Markdown; missing alt text is the one that blocks a PDF/UA export, while heading levels and the title are normalised for you at export time.
No. The check runs in memory; a share link keeps only the issue summary, and only if you create one.
No. The check reads your Markdown, not the PDF. The export still has to build a tagged file with embedded fonts, and reading order, alt-text quality and colour contrast need a human or a screen reader. Your institution may also apply its own checklist, so verify the exported file before you submit.
It parses the same Markdown the exporter does, so GitHub-flavoured Markdown, Pandoc, Quarto, MyST and MDX all work. Docs-site shortcodes are read as plain text; the check focuses on headings, images, links and the document title.
The online check accepts Markdown up to 200,000 characters, or a PDF up to 25 MB and 300 pages. For anything larger, open it in the editor — the editor accepts much larger documents, and this check is meant for a quick look before export.
Start shipping better documents
Open MarkdownGo, paste your Markdown, and let the page take care of the rest.