WordPress PDF Accessibility: 6 Ways to Fix Inaccessible Documents
The cheapest accessible PDF is the one you do not publish. That sounds glib and it is the single most useful piece of advice in this category, because remediating a PDF properly is slow specialist work and republishing the same content as an HTML page is usually faster, cheaper, better for search, and accessible by default.
PDFs are explicitly in scope for the European Accessibility Act and for the DOJ rule covering US public bodies, and for most organisations with a document library they represent the largest single body of inaccessible content. This article covers when to remediate and when to replace, then compares the tools that do the remediation.
Verified August 2026. Related: our public sector guide covers the April 2026 DOJ deadline.
Why PDFs fail, and what fixing means
A PDF is a description of where marks sit on a page. It has no inherent notion of a heading, a paragraph, a table, or a reading order, which is precisely the information assistive technology needs. Tagging adds that structural layer, and an untagged PDF is essentially a picture of a document to a screen reader.
Proper remediation means tagging every structural element, setting a correct reading order, adding alternative text to images, marking table headers so cell relationships are announced, setting the document language, and adding a meaningful title. The relevant technical standard is PDF/UA, alongside WCAG and Section 508 depending on your jurisdiction.
The important consequence: a scanned document is not a PDF problem, it is an image problem. Scanned pages need OCR before any of the above is possible, and OCR output needs checking. That is why archives of scanned material are so expensive to remediate and why triage matters more than tooling.
The 6 options compared
1. Republishing as HTML
The option to consider first and the one most organisations skip. A WordPress page is accessible by default in ways a PDF never is: it reflows on a phone, respects the reader’s font size and contrast preferences, works with every assistive technology without a tagging layer, and is indexed properly by search engines.
For most documents a public body or business publishes, a policy, a guide, a set of minutes, a price list, there is no reason it is a PDF beyond habit and the fact that someone had it open in Word. Converting takes minutes with a copy and paste and a heading pass, against hours of specialist remediation. Keep PDFs for documents that genuinely need fixed pagination: forms people print, certificates, and things with legal formatting requirements. Everything else is better as a page, and the accessibility improvement is a side effect of a change worth making anyway.
- Cost: Staff time only
- Covers: Most documents, completely
- Best for: Policies, guides, minutes, anything without fixed pagination needs
- Watch out for: Redirect the old PDF URL so links keep working
2. Adobe Acrobat Pro
The tool most people already have and the baseline for remediation work. It provides a full accessibility toolkit: an accessibility checker, the reading order tool, tag editing, and alternative text handling, and it can do the entire job properly.
The honest limitation is throughput. Acrobat is time-intensive compared to automated solutions, especially at volume, and requires significant manual intervention and accessibility expertise to remediate comprehensively. Someone who knows what they are doing can fix a ten-page document in a reasonable time; someone learning will take a day and produce something that still fails validation. For a handful of important documents it is the right tool and you already own it. For an archive of thousands, it is a staffing plan rather than a solution.
- Cost: Subscription; most organisations already have it
- Covers: Complete remediation, manually
- Best for: A small number of important documents
- Watch out for: Time-intensive and needs real expertise at volume
3. CommonLook PDF
The specialist standard, working as a plugin for Adobe Acrobat and available as a desktop application or through a browser on any operating system. It is built specifically to test, repair, and generate compliance reports for accessible PDFs, and that reporting is what distinguishes it.
It is commonly recommended for US Federal use, which tells you what it is optimised for: rigorous validation against Section 508 and PDF/UA, with documentation you can hand to a compliance officer. If your obligation includes evidencing that a document conforms rather than merely believing it does, that report is the product. Pricing is not publicly published, so get a quote. For a public body with a deadline and an audit trail requirement, this is the tool the sector uses.
- Cost: Not publicly published; request a quote
- Covers: Testing, repair, and compliance reporting
- Best for: US Federal and anyone needing evidenced conformance
- Watch out for: Works alongside Acrobat rather than replacing it
4. axesPDF
The PDF/UA specialist, able to solve more than 50 of the most difficult compliance requirements for the standard and to test and validate conformance. Where CommonLook is associated with US Federal work, axesPDF is the tool most often recommended for shift-left authoring.
That distinction matters more than any feature. Shift-left means fixing the source document rather than the exported PDF, so a properly structured Word file exports to a properly tagged PDF and needs no remediation at all. For an organisation that produces documents continuously, this is the only approach that scales: train authors to use real heading styles, real tables, and alt text in the source, and the remediation queue stops growing. Remediating exports forever is a treadmill; fixing authoring is a solution.
- Cost: Not publicly published; request a quote
- Covers: PDF/UA validation and shift-left authoring workflows
- Best for: Organisations producing documents continuously
- Watch out for: Requires changing how authors work, not just what you buy
5. Document Library Pro and PDF Embedder
The WordPress delivery layer, and a component with its own accessibility responsibilities separate from the documents themselves. How you list and present PDFs affects whether people can find and use them.
Three things matter here. Link text must describe the document rather than saying “download”, and should state the format and size so nobody is surprised by a 40MB file on mobile data. Document libraries with search and filtering must have accessible controls, since a table of 400 documents with an inaccessible filter is unusable. And embedded PDF viewers frequently trap keyboard focus or fail to expose their controls, which is why offering a plain download link alongside any embed is good practice. None of this fixes an untagged PDF, and all of it determines whether a tagged one is reachable.
- Cost: Free tiers available; paid tiers for library features
- Covers: Presentation, listing, and linking
- Best for: Making documents findable and their links meaningful
- Watch out for: Embedded viewers trapping keyboard focus
6. Outsourced remediation
The realistic answer for a large legacy archive, and worth pricing rather than dismissing. Specialist services remediate documents at a per-page rate, which turns an open-ended internal project into a quotable cost.
The arithmetic usually favours it for volume. Training staff to remediate competently takes weeks, and remediation is slow even once learned, so an organisation with 2,000 documents faces a project measured in person-months. The correct approach is almost always hybrid: triage ruthlessly first so you are only paying for documents people actually use, republish what can be HTML, then outsource what genuinely must remain a remediated PDF. Paying to remediate an archive nobody opens is the most common way to waste an accessibility budget.
- Cost: Typically per page; quotable
- Covers: Volume remediation without internal capacity
- Best for: Large legacy archives after triage
- Watch out for: Triage first, or you pay to fix documents nobody reads
Comparison table
| Option | Cost | Scales to volume | Compliance report | Best for |
|---|---|---|---|---|
| Republish as HTML | Staff time | Yes | Not applicable | Most documents |
| Adobe Acrobat Pro | Existing subscription | No | Checker only | A few key documents |
| CommonLook | Quote | Moderate | Yes | Evidenced conformance |
| axesPDF | Quote | Yes, at source | PDF/UA validation | Continuous document production |
| Library and embed plugins | Free tiers available | Not applicable | No | Delivery and findability |
| Outsourced remediation | Per page | Yes | Usually included | Large legacy archives |
Triaging an archive you cannot fix entirely
- Pull download statistics first. In most libraries a small fraction of documents accounts for nearly all downloads. That fraction is your actual scope.
- Republish the top documents as HTML. Highest use, biggest accessibility gain, and a search benefit you were not getting from a PDF.
- Remediate what must stay a PDF. Forms, certificates, and anything with legal formatting requirements.
- Archive the long tail with a stated process. Publish that accessible versions are available on request, and mean it. This is a recognised approach for genuinely archival material.
- Fix authoring now. Every new document produced with real heading styles and alt text is one you never remediate. This is the only step that stops the problem growing.
- Record what you did and when. A documented, prioritised programme is defensible; a perfect archive nobody funded is not.
Related guides
Frequently asked questions
Are PDFs really in scope?
Yes, under the EAA and under the DOJ rule for US public bodies. Documents you publish are part of your digital offering, and for many organisations they are the largest inaccessible portion of it.
Does Acrobat’s accessibility checker mean it is compliant?
It means the automated checks passed. Reading order, meaningful alt text, and correct table header associations still need human verification, exactly as with web content.
What about scanned documents?
They need OCR before anything else, and OCR output needs checking because errors in a legal or technical document matter. Scanned archives are the most expensive category, which is another argument for triaging by actual use.
Can I just link to the PDF and say it is unavailable in other formats?
Not as a general policy. A documented process for providing accessible versions on request is a recognised approach for genuinely archival material, and it does not excuse current documents people need to use.
Is it faster to fix the Word file or the PDF?
The Word file, almost always, if you still have it. Applying real heading styles and alt text in the source and re-exporting is far quicker than tagging an export, which is the whole shift-left argument.
Should we stop publishing PDFs entirely?
Not entirely, but the default should invert. Publish as a web page unless there is a specific reason for fixed pagination. That single policy change removes most future remediation work.
The verdict
First, republish as HTML. For most documents this is faster than remediation, cheaper, better for search, and accessible without a tagging layer. It is the answer people skip because it sounds like more work than it is.
For the documents that must stay PDFs: Adobe Acrobat Pro if there are a handful, since you already have it. CommonLook where you need evidenced conformance, particularly US Federal.
If you produce documents continuously: axesPDF and a shift-left authoring change. Fixing the source stops the remediation queue growing, and nothing else does.
And triage before you spend anything. A small share of documents accounts for nearly all downloads. Paying to remediate an archive nobody opens is the most reliable way to exhaust an accessibility budget without helping anyone.
Recent Posts
- Cookie Consent Plugins for WooCommerce Stores
- Best GDPR Data Request and Erasure Plugins for WordPress
- WordPress Consent Plugins That Actually Block Scripts Before Consent
- Best WordPress Cookie Consent Plugins With Google Consent Mode v2
- WordPress PDF Accessibility: 6 Ways to Fix Inaccessible Documents
