PDF/UA compliance guide: Requirements, standards, and best practices

PDF/UA is the technical standard most teams mean when they say they need an accessible PDF. It defines how a PDF should expose structure, text, navigation, and other semantics so assistive technology can interpret the document reliably.

That matters, but it’s only part of the job. Most organizations still need a broader workflow for source authoring, conversion, remediation, validation, and accessible delivery in the browser or application user interface (UI).

If you need the cross-product overview first, start with our PDF accessibility solution hub. If you need server-side remediation right away, see the PDF/UA auto-tagging API.

What PDF/UA is

PDF/UA stands for PDF/Universal Accessibility. It’s an ISO standard for accessible PDF documents. In practical terms, PDF/UA requires a document to contain enough semantic information for screen readers and other assistive tools to understand:

• What the document title is
• What content is a heading, paragraph, list, or table
• What order the content should be read in
• What images mean, and when they’re meaningful
• How form fields should be labeled and announced
• How navigation elements such as bookmarks and language metadata should behave

A visually correct PDF isn’t automatically an accessible PDF. Two files can look identical on screen, while one is unusable with assistive technology because it lacks tags, reading order, alternate text, or proper table structure.

What PDF/UA is not

PDF/UA is not a general legal safe harbor; it’s a technical standard. A file can be closer to PDF/UA and still fail a real accessibility review if the source content is poor, the tags are wrong, or the surrounding experience is inaccessible.

PDF/UA is also not the same thing as WCAG. WCAG covers digital accessibility more broadly. PDF/UA focuses on the document format itself.

How PDF/UA relates to WCAG, Section 508, ADA, and the EU Accessibility Act

These terms are often mixed together. They overlap, but they aren’t interchangeable.

The practical takeaway is simple: PDF/UA helps with the document layer, while WCAG and regulatory frameworks shape the broader accessibility expectations around that layer.

Don’t present PDF/UA output as an automatic statement of legal compliance. Instead, treat it as an important technical control inside a larger review process.

Core technical requirements for accessible PDFs

The exact checks depend on the file and the validator, but most remediation work comes back to the same set of requirements.

Tagged structure

A usable PDF needs a meaningful tag tree. Headings, paragraphs, lists, tables, figures, and artifacts should be identified correctly. Without that structure, assistive technology is forced to guess.

Logical reading order

The content must be read in a sensible order. Multicolumn layouts, sidebars, headers, footers, callouts, and floating elements often break reading order after conversion.

Language and metadata

The document should define the primary language and include basic metadata, such as a title. Screen readers use language metadata to choose pronunciation and reading rules.

Alternate text for meaningful figures

Images, charts, diagrams, and icons that carry meaning need alternate text. Decorative assets should be treated as artifacts and not announced as meaningful content.

Accessible tables

Tables need clear row and column structure. Header cells, merged cells, nested tables, and visual-only layouts commonly fail here.

Accessible forms

Form fields need labels, predictable focus order, helpful instructions, and sensible tab behavior. A form that looks correct can still be unusable from a keyboard or screen reader.

Searchable text

If the source is scanned or image-based, OCR is usually required before remediation can succeed. A tagged image of text is still a poor accessibility outcome if the underlying text layer is missing or inaccurate.

Consistent navigation

Bookmarks, links, headings, and other navigation structures should reflect the document hierarchy. This improves both usability and auditability.

Common PDF accessibility failure patterns

Most broken PDFs aren’t broken for a single reason. They usually fail in clusters.

Untagged or badly tagged conversion output

A converter may preserve layout but produce weak tags. Common symptoms include every line tagged as a paragraph, headings flattened into plain text, or lists turned into disconnected text blocks.

Incorrect reading order after layout-heavy design

Marketing brochures, annual reports, statements, and forms often use visual composition that doesn’t map cleanly to a linear reading order.

Tables built for appearance, not semantics

Merged cells, empty spacer columns, and nested tables create heavy remediation work. If a table is important enough to read, it needs real structure.

Missing or low-value alternate text

Teams often add alt text late, at scale, or from templates. The result is vague text such as “image” or “chart.” This satisfies a field but not the user.

OCR that creates noisy text

Scanned PDFs can look fine but carry poor OCR output. If the text layer is wrong, tagging the file doesn’t fix the reading experience.

Form fields without labels or instructions

Unlabeled inputs, duplicate field names, and poor tab order are common blockers in applications, claims, onboarding, and government workflows.

Inaccessible delivery layer

A technically better PDF can still be hard to use if the viewer experience isn’t accessible. For browser-based delivery, the viewing layer matters too. Refer to our SDK accessibility overview for the viewer side of the problem.

A practical remediation workflow

Teams get better results when they stop treating accessibility as a one-time cleanup project. A repeatable workflow usually looks like this:

1. Start with the best source content possible. Good heading structure, real tables, descriptive link text, and meaningful image descriptions reduce remediation later.
2. Convert or generate the document with structure in mind. If you control generation, build accessibility into the template or output process early.
3. Run OCR when the source is scanned or image-based. Without searchable text, downstream accessibility work is weaker and more expensive.
4. Apply tagging and remediation. This is where teams fix structure, reading order, tables, alt text, forms, and metadata.
5. Validate with automated tools. Use validators to catch missing tags, language settings, table issues, and structural errors.
6. Perform manual review. Automated checks don’t tell you whether the reading order makes sense or whether alternate text is useful.
7. Publish through an accessible delivery path. Browser and application accessibility still matter after the PDF itself is improved.

For high-volume operations, manual remediation alone doesn’t scale. This is where server-side processing helps. The PDF/UA auto-tagging API is designed for pipelines that need OCR, conversion, and repeatable PDF/UA output in the same workflow.

How to test PDF accessibility

A credible accessibility process combines automated validation with manual review.

Automated checks

Automated tools are useful for finding:

• Missing or incomplete tags
• Absent document titles or language metadata
• Table structure issues
• Common form field problems
• Missing alternate text markers
• Obvious reading-order or artifact issues

These checks are a fast first pass. They aren’t the whole answer.

Manual checks

Manual review is still needed for questions such as:

• Does the reading order make sense?
• Do heading levels reflect the real structure of the document?
• Is alternate text useful, not just present?
• Are tables understandable when read by assistive technology?
• Can a user complete the form with a keyboard and screen reader?
• Are repeated headers, footers, and decorative elements announced correctly or ignored as intended?

Delivery checks

If the PDF is consumed inside a product, test the viewing layer too. Accessible PDFs are easier to use when the surrounding application supports keyboard access, focus management, labeling, and screen reader compatibility. Our PDF accessibility solution hub explains how the document layer and UI layer fit together.

Where Nutrient fits

Nutrient covers different parts of the accessible PDF workflow, depending on where the bottleneck sits.

1. Accessible delivery in web applications

Use Nutrient Web SDK accessibility features when the main need is accessible document viewing and interaction in the browser. This addresses the experience around the PDF, including keyboard access, semantic rendering, and screen reader support.

2. Server-side remediation and PDF/UA output

Use the PDF/UA auto-tagging API when you need to process high volumes of existing or converted documents. This is the right fit when accessibility work has to move into backend jobs instead of manual desktop steps.

3. Generation and backend document workflows

Use Document Engine when your team controls document generation, server-side conversion, or broader backend orchestration. This is useful when accessibility has to become part of a larger document pipeline.

4. Conversion-heavy document operations

Use PDF/UA conversion workflows when accessibility sits inside a broader pipeline that already handles Office files, OCR, scanned documents, HTML, or existing PDFs.

A common pattern is to combine these pieces:

• Create or convert the document
• Run OCR where needed
• Export or remediate toward PDF/UA
• Validate the output
• Deliver the result through an accessible product experience

PDF/UA compliance checklist for implementation teams

Use this checklist as a working review, not a one-time certification exercise.

Source and authoring

• Use real headings instead of visual styling alone
• Use real lists and tables instead of layout hacks
• Write descriptive link text
• Prepare useful alternate text for meaningful figures
• Keep form labels and instructions explicit

Conversion and generation

• Preserve document structure where possible
• Run OCR on scanned input before accessibility remediation
• Check that tags map to the intended semantics
• Verify language metadata and document title
• Review reading order after conversion

Remediation and QA

• Validate tags, reading order, and table structure
• Review alternate text quality
• Test keyboard-only form completion
• Test with screen reader workflows that match the real audience
• Recheck the document after substantial edits or reexport

Delivery

• Verify the application viewer is accessible
• Avoid introducing inaccessible download or signoff steps around the document
• Keep accessibility review inside release and content operations, not outside them

If you want a shorter operational version of this review, download the PDF accessibility checklist.

Frequently misunderstood points

These questions come up regularly when teams start working on PDF accessibility.

“If my PDF passes an automated checker, am I done?”

No. Automated validation helps, but it doesn’t replace human review. A file can pass many machine checks and still be confusing in practice.

“If I export to PDF/UA once, will future edits stay accessible?”

Not necessarily. New source content, new templates, conversion changes, or later manual edits can introduce regressions.

“Do I need both document accessibility and viewer accessibility?”

In many products, yes. The PDF can be improved while the application UI still blocks keyboard or screen reader users.

“Can auto-tagging replace all manual work?”

No. Auto-tagging reduces manual effort and improves consistency, especially at scale. It doesn’t replace source quality, manual review, or judgment for complex layouts and specialized content.

Next steps

If your main problem is understanding the full product path, go to the PDF accessibility solution hub.

If you already know you need backend remediation, start with the PDF/UA auto-tagging API.

If your biggest issue is accessible browser delivery, review the Web SDK accessibility page.