Reviewed September 30, 2026
Accessible PDFs
A PDF exported from a slide or a scan is often a picture of text. Screen readers cannot read it until the file has a tag tree, a logical reading order, and a document title. If that file is how people get the information, it has to meet the same WCAG duties as the web page that links to it.
What is the minimum bar before you publish?
- Real text, not a scan, unless you also provide an accessible HTML or tagged equivalent.
- Tags for headings, lists, tables, and figures, in the order a person would read them.
- Figures with alternative text, and decorative artifacts marked as artifacts so they are skipped.
- Tables with header cells marked as headers, and a reading order that does not jump across columns.
- A title set in the document properties, not only in the filename, and a language set on the document.
- Form fields with tooltips or labels, and a visible tab order that matches the visual order.
Why do scanned PDFs fail?
A scanner produces an image. Optical character recognition can add a text layer, but the layer is often untagged, in the wrong order, or full of errors in names and numbers. A screen reader then reads nonsense, or nothing. For a form people must fill in, an untagged scan also has no fields.
If the source is a Word, Google Docs, or InDesign file, export tags from that source instead of printing to PDF and hoping. Fix heading styles in the source. A document that uses bold text instead of heading styles exports as a pile of paragraphs.
When is HTML the better format?
Policies, agendas, bills, and instructions that you update belong on a web page. HTML is easier to resize, search, and maintain, and it does not require a separate tag tree. A PDF is reasonable for a fixed artifact such as a signed record or a print layout that must stay paginated. If you must offer a PDF, link the HTML version next to it so nobody is forced into the file.
How do you check a PDF?
Open the tags panel and walk the tree in reading order. Check that headings nest, tables have headers, and figures have alternative text. Then listen to the first two pages in a screen reader. PDF techniques are listed in the W3C WCAG PDF techniques. State and local governments that publish the only copy of a service as a PDF need that file to meet WCAG 2.1 Level AA under the Title II web rule.
Questions and answers
What makes a PDF accessible?
The file needs real text, a tag tree in a logical reading order, a document title, alternative text on figures, header cells on tables, and labels on form fields. A scan that is only a picture of text is not accessible until an equivalent is provided.
Does WCAG apply to PDFs?
Yes. WCAG success criterion 1.3.1 (Info and Relationships, Level A) applies to the information and relationships in the file the same way it applies to a web page. If the PDF is the only copy of the information, it has to meet the criteria that apply.
Should policies be published as PDF or HTML?
Policies, bills, and instructions that you update belong on a web page. A PDF is reasonable for a fixed artifact such as a signed record or a print layout. If you must offer a PDF, link an HTML version next to it so nobody is forced into the file.
Is a PDF/UA file the same as a WCAG-conformant file?
PDF/UA is a separate ISO standard for accessible PDF. A file that meets PDF/UA is usually in good shape for the WCAG techniques that apply to PDF, but you still check the content: reading order, table headers, and whether the text alternative matches the figure.
Related guides
Ask for a review of a specific page
Send the URL and the template you care about. The reply names the highest-risk WCAG failures and the order to fix them.
Request an audit