What is a tagged PDF? Tags, structure, and how to tag a PDF

A tagged PDF is a PDF that includes a logical structure layer on top of its visual content. Each element on the page is labeled with a tag that states what it is: a heading, a paragraph, a list item, a table cell, or a figure. The tags sit in a structure tree, which also defines the order in which the content should be read.

An untagged PDF has none of that. To software, it’s a collection of text fragments and images positioned on a page, and a screen reader has to infer the structure from that layout. Correct tagging gives assistive technology explicit information about structure and reading order, which is why PDF/UA, the ISO standard for accessible PDFs, requires it.

Tagged vs. untagged PDF

The table describes a correctly tagged PDF. Tags can be missing or incorrect, even when a file is marked as tagged.

Two PDFs can look identical onscreen while one is tagged and the other isn’t. The difference only shows up when software tries to interpret the content.

Searchable PDF vs. tagged PDF

A searchable PDF contains text you can find or select. A tagged PDF also describes the text’s role and its place in the document’s reading order. Being able to search for a heading doesn’t mean a screen reader can recognize it as a heading.

For a scanned PDF, optical character recognition (OCR) turns images of words into text. OCR alone doesn’t establish the tag structure. Some tools combine recognition and tagging, but you still need to check both the recognized text and the tags, as W3C’s OCR guidance(opens in a new tab) explains.

Example of a tagged PDF structure

A tag tree mirrors the logical structure of the document, not its visual layout. A simple report with a heading, an introduction, a list, a table, and a chart might be tagged like this:

Document
├── H1          "Quarterly report"
├── P           "Revenue grew across all regions."
├── L
│   ├── LI
│   │   ├── Lbl    "•"
│   │   └── LBody  "EMEA"
│   └── LI
│       ├── Lbl    "•"
│       └── LBody  "AMER"
├── Table
│   ├── TR
│   │   ├── TH  "Region"
│   │   └── TH  "Revenue"
│   └── TR
│       ├── TD  "EMEA"
│       └── TD  "$1.2M"
└── Figure      Alt: "Bar chart showing revenue growth in every region"

A screen reader follows this tree from top to bottom. It announces the heading as a heading, the list as a list with two items, and the table with its column headers. The figure is read through its alternative text. Page numbers, running headers, and decorative lines are marked as artifacts outside the tree, so assistive technology skips them.

Common PDF tags and what they mean

PDF defines a set of standard structure types. These are the ones most documents use:

Tags describe meaning, not appearance. A line of large, bold text should be tagged H2 only if it really is a section heading.

How to check whether a PDF is tagged

Use a PDF accessibility checker to inspect the file:

1. Open the PDF in the checker and run its accessibility checks.
2. Review the results for missing tags, untagged content, and structural errors.
3. In a tool that exposes the tag tree, expand it and compare the tags with the page. Check that headings, lists, tables, and figures have appropriate roles and follow a logical reading order.

For developers inspecting the PDF directly, the document catalog contains a MarkInfo dictionary with Marked set to true and a StructTreeRoot entry for the structure tree. These entries identify declared tagging; their presence alone doesn’t establish that the content is correctly tagged.

A file can have a structure tree with the wrong tag types, a broken reading order, or figures without descriptions. Review those details even if the checker reports that tags exist.

For a concrete example, see the tagged PDF with a missing language setting. Its heading and paragraph tags were present, but veraPDF reported language failures for page text and metadata. Setting the document language resolved those automated failures without changing the tags. The downloadable PDFs and reports show why tagging and validation are separate checks; passing validation still leaves manual accessibility review.

How to tag a PDF

There are three common routes to a tagged PDF. The right one depends on whether the source document exists and how many files need processing.

Export a tagged PDF from the source

When a well-structured source document exists (one with real headings, lists, and tables rather than formatting that only looks like them), export a tagged PDF from the authoring application. This lets you correct structure in the source document and reuse those corrections in future exports.

For example, in Word for Microsoft 365 on Windows:

1. Run Word’s Accessibility Checker and address the issues it identifies.
2. Open File > Save As or Save a Copy. Choose Browse and select PDF as the file type.
3. Open Options and select Document structure tags for accessibility.
4. Select OK and save the PDF. Open the exported file in a PDF accessibility checker to review its tags.

These steps follow Microsoft’s accessible PDF export instructions(opens in a new tab), which also cover macOS and the web. Exporting with tags doesn’t replace checking the resulting PDF.

Auto-tag existing PDFs

For PDFs that already exist without tags, auto-tagging detects document elements and adds the structure tree. The Nutrient Accessibility API auto-tags PDFs with headings, paragraphs, lists, and tables, and improves reading order. The auto-tagging guide covers sending a file to the /accessibility/autotag endpoint and receiving tagged output.

Teams working in .NET can also auto-tag and convert PDFs to PDF/UA with the .NET SDK.

Review the tags that auto-tagging assigns. Alternative text quality, complex layouts, and meaning that depends on visual context still need human review, as Nutrient’s introduction to PDF/UA and auto-tagging explains.

Generate tagged PDFs from code

When an application creates PDFs, it can build the structure tree directly. With Nutrient .NET SDK, the process is:

1. Set the document title and language with SetTitle and SetLanguage.
2. Get the root tag with GetTagRootID.
3. Create tags such as Sect and Figure under the root with NewTag.
4. Add content inside each tag as marked content.
5. Set alternative text for figures with SetTagAlternateDescription.

The tagged PDF guide includes a complete C# and VB.NET example.

Whichever route is used, check the result with a validator and a screen reader before publishing. For files that need more than tags — OCR, image descriptions, or corrected reading order — see the PDF remediation guide.

FAQ