An image in a PDF is, to a screen reader, nothing at all — unless somebody attached a text description to it. Alt text is that description, and adding it is one of the cheapest accessibility improvements available: a few seconds per image, no design change, and the difference between a chart that conveys its finding to every reader and one that is a silent gap in the document. This covers how alt text works in PDF specifically, how to add it in the common tools, and how to write text that is actually useful.
How alt text works in a PDF
Unlike HTML, where alt is an attribute on the element, PDF stores the description in the structure tree — the tagged-document hierarchy that describes the logical content of the page. An image becomes a /Figure structure element, and that element carries an /Alt string.
Two consequences follow:
The document must be tagged. An untagged PDF has no structure tree, so there is nowhere for alt text to live. If your file is untagged, tagging is the prerequisite step, not an optional extra — see tagged PDF vs untagged PDF and accessibility tags in PDF explained.
Decorative images need marking too. An image with no informational content — a divider rule, a background texture, a decorative flourish — should be marked as an artifact rather than given empty alt text. Artifacts are excluded from the reading order entirely, which is exactly right; a screen reader should not announce them at all. This is the PDF equivalent of alt="" in HTML, and it matters more here because an unmarked image is announced as an unlabelled figure, which interrupts reading for no reason.
There is also /ActualText, a related field with a different job: it replaces the content for a reader rather than describing it. Use it when an image is literally text — a logo containing the company name, a decorative drop cap, a scanned equation — so the reader receives the words. Alt describes; ActualText substitutes.
Adding alt text in Acrobat Pro
The reliable route, and the one most accessibility work uses.
For a whole document at once, use the Set Alternate Text tool:
- Tools → Accessibility → Set Alternate Text.
- Acrobat steps through every figure in the document one at a time, showing it and offering a text box.
- Type the description, or tick Decorative figure to mark it as an artifact.
- Use the arrows to move through all the images.
This is the fastest path for a document with a dozen images, and its explicit Decorative checkbox handles the artifact case properly.
For a single image, use the Tags panel:
- View → Show/Hide → Navigation Panes → Tags.
- Find the
<Figure>element (the Find Tag from Selection option locates it from a selection on the page). - Right-click → Properties → Alternate Text.
The Tags panel is also where you fix the more common underlying problem: an image that has not been tagged as a Figure at all, or has been swept into a <P> with surrounding text.
Check your work with Tools → Accessibility → Accessibility Check (Full Check). The report flags every figure missing alternate text, which is the fastest audit available.
Adding it before you export
Far better than retrofitting: put the alt text in the source document, where it will carry through the PDF export automatically.
Microsoft Word: right-click an image → View Alt Text, type the description, or tick Mark as decorative. Export via File → Save as Adobe PDF or Save As → PDF with Document structure tags for accessibility enabled in the options. Word's PDF export carries alt text correctly, which makes Word one of the more accessible authoring routes despite its other flaws.
PowerPoint: identical mechanism, same export caveat.
LibreOffice Writer: right-click the image → Properties → Options → Alternative (Text only). Export as PDF with Tagged PDF ticked in the export dialog.
Google Docs: right-click the image → Alt text. Download as PDF carries it.
Adobe InDesign: Object → Object Export Options → Alt Text, where you can also set the source to the image's XMP description so alt text lives with the asset. InDesign's export produces properly tagged figures when Create Tagged PDF is enabled.
LaTeX: the pdfcomment or tagpdf packages, or \pdftooltip{}{} from pdfcomment for simple cases. Accessible LaTeX output remains harder than it should be.
The rule of thumb: fixing accessibility in the source costs minutes; fixing it in the exported PDF costs hours, and has to be redone every time the document is regenerated.
Writing alt text that helps
Adding a description is half the job. Writing a good one is the other half, and it is where most documents fail even when the field is populated.
Describe the information, not the picture. For a bar chart, "Bar chart showing quarterly revenue" is nearly useless — it tells the reader a chart exists and nothing about what it says. "Quarterly revenue 2026: Q1 £1.2m, Q2 £1.4m, Q3 £1.1m, Q4 £1.9m, with the Q3 dip attributed to the supply delay" conveys the same thing the sighted reader gets.
Match the length to the role. A logo needs three words. A photograph illustrating a point needs one sentence. A complex diagram needs a short alt text plus a full description in the body text or an appendix — alt text is announced as a single unbroken block, and four hundred words of it is punishing to listen to.
Do not start with "image of" or "picture of". The screen reader already announces it as a figure. The prefix wastes the listener's time on every single image.
Do not duplicate the caption. If the caption already states what the figure shows, the alt text should add what the caption cannot — the actual values, the trend, the visual detail being pointed at.
Include text that appears in the image. Any words rendered inside a graphic are invisible to assistive technology. If a diagram is labelled, those labels belong in the description.
For data-heavy figures, provide the data. The genuinely accessible answer to a complex chart is a real table alongside it — see making accessible PDF tables. Alt text summarises; the table carries the numbers.
Mark decoration as decoration. If the honest description is "a stock photograph of people in a meeting, illustrating nothing", it is decorative. Mark it as an artifact and let the reader past.
Scanned documents
A scanned page is one large image, and alt text is the wrong tool for it entirely. Describing a page as "scan of page 4" gives the reader nothing. The correct fix is OCR, which turns the picture of text into text: how to make a PDF searchable with OCR and PDF OCR explained.
Where a scanned page contains a genuine image — a photograph within a scanned article — that image needs alt text once the page has been OCR'd and tagged.
Auditing an existing document
To find what is missing:
- Acrobat's Full Check lists every figure without alternate text, with a click-through to each.
- PAC (PDF Accessibility Checker), free for Windows, gives a more rigorous PDF/UA-oriented report and is the standard second opinion.
veraPDFvalidates against PDF/UA formally, which is what you want if compliance is being asserted rather than merely aimed at.
Volume matters here. A report with three hundred images will not be fixed by hand; the practical approach is to fix the source template, regenerate, and reserve manual work for one-off documents. See PDF/UA accessibility standard explained and WCAG 2 for PDF accessibility for what compliance actually requires.
Summary
Alt text lives in the tagged structure tree, so the document must be tagged before there is anywhere to put it. Add it in the source application — Word, InDesign, LibreOffice — and it survives the export; retrofit it in Acrobat's Set Alternate Text tool when you cannot. Describe what the image tells the reader rather than what it looks like, skip the "image of" prefix, mark decorative graphics as artifacts, and give complex figures a real data table rather than a paragraph of alt text nobody can follow by ear.
