How to Make a PDF Accessible

Path 1: Build from a structured source

1. Use heading styles, not manual formatting

   Make a heading a Heading 1, Heading 2, or Heading 3 in the authoring tool; changing the font size and weight is not enough. In LaTeX, use commands such as \section{} and \subsection{}. Keep levels in a sensible order so users can understand the outline and move through it quickly.

2. Add alt text to every meaningful image

   Describe the information the image contributes in this context. For a chart, state the main trend or result instead of writing only “chart.” Put long descriptions near complex diagrams and reference them in the alt text. Mark purely decorative images so assistive technology skips them.

3. Structure tables with real cells and headers

   Do not align text with tabs or draw a table as an image. Use the authoring tool's table object, identify the header row, avoid merged cells where possible, and repeat header rows across pages. See the guide to accessible tables for Word, LaTeX, Google Docs, and Acrobat.

4. Set the document language and title

   Choose the main language in the source document and mark passages that switch languages. Add a concise title in document properties; a file name such as “BIO101-final-v3.pdf” is not a useful title.

5. Run the source accessibility checker

   Fix missing alt text, skipped heading levels, unclear link text, low contrast, and table warnings before export. Source-level repairs are faster and usually produce a cleaner tag tree than patching the PDF later.

6. Export with structure tags enabled

   Use Save As or Export to PDF and keep the accessibility or document-structure option turned on. Never use Print to PDF for the final accessible version: printing commonly flattens or drops the logical structure you just created.

Path 2: Remediate an existing PDF

Text-based PDFs

If text is selectable, use an automatic tagging tool to build a first draft of the structure. Then inspect the Tags panel. Establish one Document root, correct the heading hierarchy, tag paragraphs and lists, mark decoration as artifacts, repair links, add figure alt text, and place every element in the intended reading order. Automatic tagging is a starting point, not a finished result.

Scanned PDFs

Run OCR first and compare the recognized text with the page. Fix errors before tagging; otherwise assistive technology receives the wrong content in a clean-looking structure. After OCR, create the tag tree, define reading order, describe figures, rebuild tables, and add bookmarks for long documents.

For large, technical, or mixed-quality collections, a document remediation service can be more reliable than repeating this process by hand. See Underleaf PDF remediation for scanned and handwritten STEM work.

How to verify the result

• Run Adobe Acrobat Pro's full accessibility check and inspect every item that requires manual review.
• Run PAC when PDF/UA conformance is a goal, then fix errors in the source or remediation tool.
• Review the full tag tree and reading order page by page, including footnotes, captions, sidebars, links, and form controls.
• Test representative pages with NVDA on Windows or VoiceOver on macOS. Navigate by headings, links, lists, and tables.
• Zoom and reflow the document, verify contrast, and check that information does not depend on color alone.

If you are new to the structure panel, first read What Is a Tagged PDF? and the broader accessible PDF guide.

Quick reference by source format