Guide
How to check if a PDF is accessible, free
Four ways to check a PDF against PDF/UA and WCAG without paying for anything, what each one catches, and the failures none of them can see.
About 6 minutes · quick
The short answer
Run the file through two checkers, not one, and then listen to it. Every free tool has a blind spot, and the overlap is what you can trust.
| Tool | Cost | Platform | What it is good at |
|---|---|---|---|
| veraPDF | Free, open source | Any (Java) | The reference PDF/UA validator. Cite this one. |
| PAC 2024 | Free | Windows | Screen-reader preview and structure inspection |
| Acrobat Pro checker | Paid | Win/Mac | Fixing, not finding. Weak checker, strong editor. |
| Taggart | Free, unlimited | Browser | Matterhorn conditions plus right-to-left analysis |
1. veraPDF — the one to cite
veraPDF is the PDF Association’s reference implementation. When a compliance report needs a citable validation result, this is what it should say.
# macOS / Linux, after installing the greenfield build
verapdf --flavour ua1 --format text yourfile.pdf
It gives you a pass or fail per clause of ISO 14289-1, with the object ids. What it does not give you is any sense of what to do about it — the output is a conformance verdict written for a machine.
Its blind spot is everything requiring judgement, which it will tell you honestly: veraPDF reports what it can decide and stays silent on the rest.
2. PAC 2024 — the one that shows you the damage
PAC (PDF Accessibility Checker) is free, from Access for All, and its value is not the checker — it is the screen-reader preview. It renders the document as a tag tree with the text a reader would announce, in the order it would announce it.
Open your file, run the check, then open the preview and read it top to bottom. Reading-order failures are obvious in about four seconds in this view and nearly invisible everywhere else.
Its blind spot is that it is Windows-only, which for a designer on a Mac means a virtual machine or nothing.
3. Acrobat Pro’s Accessibility Checker
If you already have Acrobat Pro: All tools → Prepare for accessibility → Check for accessibility.
Be aware of what this tool is. It reports 32 items, not 136, and several of its checks are lenient in ways that matter. Its “Tagged PDF” check passes if a structure tree exists, regardless of whether the tree corresponds to the content. Its “Logical Reading Order” check does not check reading order — it prints an instruction telling you to check it manually, and then marks the item as needing manual verification, which many people read as a pass.
Use it for repair. Do not use it as your evidence.
4. Taggart’s validator
Free, unlimited, no account, no page cap, no watermark: open it and drop a file in.
It reports every machine-checkable Matterhorn failure condition with the page, the structure path and the object id, and it adds the right-to-left analysis no other tool performs — the visual-order detection that finds Hebrew and Arabic text stored pre-reversed.
Checking is commoditised and we price it accordingly. Remediation is what a plan pays for.
The ten-second test that needs no tool at all
Open the PDF in any reader. Press Ctrl/Cmd+A, then Ctrl/Cmd+C, and paste into a plain-text editor.
What you get is approximately what a screen reader gets. If the paste is empty, the pages are images — no amount of tagging will help until it is OCRed. If the paste is jumbled, the reading order is wrong. If a Hebrew or Arabic paste comes out mirrored, the text is stored in visual order.
Three of the most expensive defects in this field, found with two keystrokes.
What no free checker can tell you
This is the part worth internalising before you build a process around any of these tools.
The Matterhorn Protocol enumerates 136 failure conditions. Roughly 87 can be decided by software. 47 require a human, and they include the ones that decide whether the document actually works:
- Is the reading order the intended order? A checker sees an order, not the right one.
- Does the alt text describe the image, or does it say “image”?
- Is that a data table or a layout table?
- Is that heading a heading, or a line of bold text that looks like one?
- Is the document language the language it is actually written in?
Two conditions have no test at all. A tool that reports “compliant” has told you about at most 64% of the protocol, and it has told you nothing about the 36% a user would notice first.
That is why every serious remediation workflow ends with a person, and why the deliverable is a report with decisions in it rather than a green tick.
Frequently asked questions
Which free checker is the most authoritative?
veraPDF is the reference implementation for PDF/UA validation and is the one to cite in a report. PAC is the more useful day-to-day tool because its screen-reader preview shows you what the tags actually produce. They disagree occasionally, and where they do, veraPDF's reading of the standard is the one that holds up.
Acrobat says my PDF passed. Is it accessible?
It means no machine-detectable failure was found. Of the Matterhorn Protocol's 136 failure conditions, 47 require human judgement and cannot be decided by any checker — including whether the reading order is the right one and whether alt text describes the image. A pass is the beginning of the check, not the end of it.
Do I need Acrobat Pro to check a PDF?
No. veraPDF, PAC and Taggart's validator are all free and none of them require Acrobat. You need Acrobat Pro to repair a file by hand, not to find out what is wrong with it.
Is there a checker that works on macOS?
veraPDF runs anywhere Java runs, including Apple Silicon. PAC is Windows-only. Taggart's validator runs in the browser, so the platform question does not arise — which is why the Mac half of this market has been so badly served.
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.
- 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.
- 11-001 Natural language for text in page content cannot be determined No /Lang is in effect for some text: neither on the catalog, nor on an enclosing structure element, nor on a marked-content sequence.
- TG-RTL-001 Right-to-left text is stored in visual order Older Israeli government, legal and DTP systems lay out Hebrew by reversing the character order and then rendering left-to-right. The page looks correct, but the stored text is backwards. Extracting it gives you the sentence mirrored, and the bidi algorithm then reverses it a second time.
Check your own file. Taggart's validator is free and unlimited — no page cap, no watermark, no account.
Last updated .