How to Create Accessible PDFs With Proper Reading Order, Tags, Links, and Alternative Text

Create the PDF from a clean source file, then verify tags, reading order, links, and alternative text before you publish it. That single workflow prevents most accessibility failures and saves painful repair time later.

Start with structure, not decoration

A PDF is accessible when assistive technology can understand it. That means the file cannot be just a visual copy of a page. It needs hidden structure called tags. Tags tell a screen reader, braille display, or text extraction tool what each item is: a heading, paragraph, list, table, image, link, or form field.

If you only make text big and bold, it may look like a heading, but it may not act like one. Use built-in styles such as Heading 1, Heading 2, and Body Text. Use real bullets and numbered lists. Do not fake spacing with extra tabs or empty lines. That stuff often turns into PDF clutter.

Build the source document correctly

The easiest PDF to fix is the one you do not break in the first place. Start with an accessible source document. This could be a Word file, a Google Docs file, a layout file, or an HTML page. The same rules apply.

• Use heading styles. Keep one main title, then nest sections in order.
• Use real lists. Do not type hyphens by hand and call it a list.
• Use simple tables. Mark header rows and avoid split or merged cells when possible.
• Add alt text before export. It usually travels into the PDF better that way.
• Name links clearly. Link text should say where the link goes or what it does.
• Set the document language. This helps pronunciation in screen readers.

Honestly, it feels like some PDF tools punish you for caring about accessibility. A tiny layout change can shift the reading order in three places. Expect to spend a few extra minutes checking the final file, even when the source document is great.

Set the reading order so content makes sense

Reading order is the order in which assistive technology reads the page. Sighted readers can scan columns, sidebars, captions, and pull quotes. Screen reader users hear one item at a time. If the order is wrong, the page becomes a puzzle.

A bad reading order might read the footer first, then a chart caption, then the second column, then the page title. That is not a small bug. It can make the PDF unusable.

Check reading order page by page. In Acrobat, use the Reading Order tool and the Tags panel. Other PDF editors have similar panels. Look for these common problems:

1. Headers and footers read before body content. Mark repeated decorative items as artifacts when needed.
2. Multi-column text reads across columns. It should finish the left column before moving to the right column.
3. Captions separate from images. Keep each caption near the image it describes.
4. Sidebars interrupt main text. Place them where they make sense in the reading flow.
5. Page numbers announced at awkward times. Artifact them if they add noise.

Read the page out loud in the tag order if you are unsure. If it sounds confusing to you, it will likely be worse through assistive technology.

Use PDF tags the right way

Tags are the backbone of an accessible PDF. A clean tag tree usually follows the visible structure of the document. It should not be packed with random spans, empty tags, or unlabeled containers.

Use <H1> for the main title, <H2> for major sections, and lower heading levels only when the content truly needs them. Use <P> for paragraphs. Use <L>, <LI>, <Lbl>, and <LBody> for lists. Tables should have table tags, header cells, and data cells.

Do not tag decorative lines, background graphics, or repeated borders as meaningful content. Mark them as artifacts. Screen reader users do not need to hear “blue rectangle” six times on a page.

Write links that tell users what will happen

Accessible links need two things: a working target and clear link text. Avoid vague labels such as click here, learn more, or read this. They make no sense when a screen reader lists all links on a page.

Better link text sounds like this:

• Download the 2026 tax form
• Read the employee handbook
• Email the accessibility team
• Open the registration portal

If the link opens a file, say so. If it starts an email, say so. If it opens a new site, the context should be clear. Also check that links are tagged as links, not just colored text. A blue underlined phrase with no link tag is only decoration.

Add alternative text that helps, not annoys

Alternative text describes meaningful images for people who cannot see them. Good alt text is short, specific, and tied to the purpose of the image. It does not need to describe every pixel.

For a photo of a doctor speaking with a patient, weak alt text would be image or doctor. Better alt text would be Doctor explaining test results to a patient. If the surrounding paragraph already explains the image, the alt text can be shorter.

Charts need more care. Do not write only bar chart. Give the key message. For example: Bar chart showing customer satisfaction rising from 68% in 2022 to 81% in 2024. If the chart is complex, include a longer description nearby in the text.

Decorative images should usually have empty alt text or be marked as artifacts. A flourish, divider, or stock background image should not interrupt the reading experience.

Make tables readable

Tables can be brutal in PDFs if they are not tagged well. Keep them simple. Use one header row where possible. Avoid nested tables. Do not use a table just to control page layout.

Each data cell should connect to the right header. For a pricing table, a screen reader should make it clear which plan and feature match each price. If a user hears “$29” with no context, the table has failed.

If the table is too complex, consider splitting it into smaller tables. You can also add a plain-language summary before it. People appreciate that. So do search engines and compliance reviewers.

Test the PDF before sharing it

An automated checker is useful, but it cannot catch everything. It may confirm that an image has alt text, but it cannot always tell whether the alt text is helpful. It may detect headings, but not whether the order makes sense.

Use this quick test plan:

1. Run an accessibility checker. Fix missing tags, missing language, empty headings, and untagged content.
2. Inspect the tag tree. Look for broken order, empty tags, and wrong tag types.
3. Use keyboard controls only. Make sure links and form fields can be reached in order.
4. Try a screen reader. Listen to at least the first two pages and any complex section.
5. Zoom to 200%. Check whether text remains readable and content is still usable.

It drives me crazy when a checker says “passed” while the first page reads like a junk drawer. Treat the checker as a filter, not a final judge.

Watch for scanned PDFs

A scanned PDF is often just a picture of text. Screen readers cannot read it unless the file has accurate OCR text and proper tags. Run optical character recognition, then proofread the result. OCR mistakes are common with small type, stamps, handwriting, and old photocopies.

After OCR, the file still needs reading order, tags, headings, links, and alt text. OCR alone does not make a PDF accessible. It only gives you text to work with.

Final checklist

• The PDF has a document title and language.
• Headings are tagged and nested correctly.
• Reading order matches the visual and logical order.
• Links have clear names and valid targets.
• Images have useful alt text or are marked decorative.
• Tables have headers and a simple structure.
• Forms have labels, instructions, and a logical tab order.
• Automated and manual tests are complete.

Accessible PDFs are not about checking a legal box. They are about letting people read, search, understand, and act without asking someone else for help. Build the structure early, test the file like a real user, and the PDF will work better for everyone.