Guide
How to mark page furniture as artifacts
Headers, footers, rules and background images should be excluded from the reading order, not given empty alt text. The distinction, and how to apply it.
About 9 minutes · moderate
The distinction
Everything drawn on a PDF page is one of two things:
Content — it is in the structure tree and a screen reader encounters it.
An artifact — it is excluded from the structure tree and a screen reader never sees it.
There is no third category and no partial state. Content that should be excluded interrupts the reader; artifacts that should be content disappear from the document.
Getting this wrong is common because the two failures look opposite and have the same cause: nobody made the decision, so the tagger guessed.
What should be an artifact
Running headers and footers. The document title repeated at the top of every page, the department name at the bottom. A sighted reader glances past these; a screen-reader user hears them twice per page transition.
Page numbers. Same reasoning, and the standard has a specific artifact type for them.
Decorative rules and borders. The line under a heading, the box around a pull-quote, the ornament on a section divider.
Background images and watermarks. A logo tinted at 8% behind the text is not information.
Printer’s marks. Crop marks, registration targets, colour bars.
Repeated design elements. The coloured band down the side of every page, the pattern in the margin.
What should not be an artifact
Anything with information in it. A logo in a header is decorative on page 7 and is the issuing organisation on page 1 — where it is often the only place the organisation is named.
A watermark that says something. “DRAFT”, “CONFIDENTIAL” and “SUPERSEDED” are content. A reader who does not know a document is a draft has been given wrong information, and this is one of the more consequential mistakes in this area.
A chart’s data. Charts are frequently drawn as vector paths, and vector paths get swept into artifacts by careless bulk operations. The chart is content and needs alt text.
A signature block. It looks like an image and it is a legal element of the document.
A decorative image that is also a link. Interactive means content, always.
The four artifact types
PDF defines four, and using the right one matters for tools that process artifacts selectively:
| Type | For |
|---|---|
Pagination | Headers, footers, page numbers |
Layout | Rules, borders, boxes, decorative separators |
Page | Printer’s marks, crop marks, colour bars |
Background | Watermarks, background images and tints |
In the content stream an artifact is a marked-content region that no structure element references:
/Artifact << /Type /Pagination /Subtype /Header /BBox [0 750 612 792] >> BDC
BT /F1 9 Tf 1 0 0 1 72 765 Tm (Annual Report 2025) Tj ET
EMCThe BDC/EMC pair wraps it; the absence of an MCID means nothing in the structure tree can point at it.
In Acrobat Pro
The Reading Order tool is the right instrument here, and it is genuinely fast:
All tools → Prepare for accessibility → Fix reading order.- Drag a marquee around the header. Choose Background/Artifact.
- Repeat for the footer and page number.
- For the whole document, use the Tags panel: right-click an existing artifact-worthy element → Change Tag to Artifact, and choose the type.
For running elements that repeat on every page, do one page and then check whether your producer emitted them identically — often the same object is reused, and marking one marks all.
The reverse operation, when something has been artifacted that should not have been: in the Reading Order tool, click Show Background/Artifact Content at the bottom of the panel. Artifacts appear with a grey overlay. Marquee the region and give it a real type.
That checkbox is worth knowing about for a different reason: it is the fastest way to audit a document that someone else remediated. Turn it on and look at what is grey. If a chart, a signature or a “DRAFT” stamp is grey, the previous pass over-artifacted.
In the authoring tool
Easier than fixing it afterwards, in every case.
Word — content in the Header and Footer areas is exported as Pagination artifacts automatically. Content in the body that looks like a header is not. Use the actual header area.
InDesign — anything on a master page that is not overridden exports as an artifact if you tick Create Tagged PDF and have not given it a role. Decorative frames should have their Object Export Options role set to Artifact explicitly rather than relying on the default.
LaTeX — \includegraphics[artifact]{ornament.pdf} with a recent graphicx.
The over-artifacting failure
There is a way of “fixing” a document that shows up regularly in outsourced remediation: artifact everything that is hard, and the checker goes quiet.
It works, in the sense that a validator cannot flag content that is not in the structure tree. It produces a document that passes every automated check and that a screen-reader user cannot read, because half of it has been declared decorative.
If you are reviewing remediation someone else delivered, this is the first thing to look for. Turn on Show Background/Artifact Content and read what is grey. Any body text there is a document that has been made to pass rather than made to work — and it is a much worse outcome than the untagged file it started as, because it now carries the appearance of having been fixed.
Frequently asked questions
Is empty alt text the same as an artifact?
No, and the difference is audible. An image with alt='' is still a Figure in the structure tree: it occupies a position in the reading order and several readers announce something for it. An artifact is excluded from the structure entirely — the reader never encounters it. For decorative content, artifact is the correct answer.
Should page numbers be artifacts?
Yes. A page number in the reading order interrupts the text on every page transition. It is page furniture, and it is exactly what the Pagination artifact type exists for. The exception is a document where the page number is referenced in the content, such as a legal filing with paragraph numbering — but that is content, not furniture.
What about a decorative image that is also a link?
Then it is not decorative. Anything interactive has to be in the structure tree with an accessible name, because a reader needs to know it is there in order to activate it. Give it alt text describing its destination, not its appearance.
Can a whole page be an artifact?
A blank separator page, a printer's colour bar, a cover flourish with no text — yes. A page with any text a reader might want is not artifact material, however decorative the layout is.
Related failure conditions
- 01-005 Content is neither marked as Artifact nor tagged as real content Every text-showing, image-painting and path-painting operator on a page must sit either inside a marked-content sequence that maps to the structure tree, or inside an artifact sequence. Content that is in neither is orphaned.
- 13-001 Graphics objects are not tagged with a Figure tag An image or vector drawing that carries meaning is not inside a Figure element. Either it is untagged, or it has been swept into a paragraph.
- 09-001 Tags are not in logical reading order The order of elements in the structure tree does not match the order a human would read the page. This is the single most common substantive defect in tagged PDFs and the one Acrobat's own tooling handles worst.
- 14-001 Headings are not tagged Text that is visually a heading — larger, bolder, isolated — is tagged as a paragraph. The document has no navigable outline.
Check your own file. Taggart's validator is free and unlimited — no page cap, no watermark, no account.
Last updated .