It's Friday afternoon and marketing just handed you a 40 page product PDF that "needs to be accessible before it goes live." Nobody tells you what that means until legal forwards a complaint, or a client's procurement team sends back a form asking for proof. You Google the fix and the first five results all point to the same answer: buy Acrobat Pro, $240 a year, a subscription nobody put in this quarter's budget.
Here's the part nobody mentions. Acrobat Pro is convenient, not required. The actual technical requirement is a properly tagged, structured PDF, and there is a full free toolchain that gets you there: export settings you already have in LibreOffice or Google Docs, a free desktop checker built specifically for this job, and a little manual cleanup. No subscription needed.
The Stat: Federal agencies are required under Section 508 to evaluate ICT purchases, including PDFs, using a VPAT/ACR before procurement. (Source: Section508.gov)
Why The Tag Structure Matters More Than The Software
A PDF isn't accessible because it was made by expensive software. It's accessible because it carries a tag tree: headings marked as headings, reading order that matches the visual layout, alt text on images, and a language attribute. Acrobat Pro has a nice interface for editing that tree. It does not create anything that a free export process can't also create, because the tags come from the document structure you build before export, not from the PDF tool itself.
That's the piece most teams miss, and it's exactly the blind spot covered in why PDFs quietly become a compliance liability behind your lead magnets. If the source document has no real heading structure, no tool, paid or free, will magically infer one for you. So step one always happens upstream, in the word processor.
This is also why buying Acrobat Pro without changing anything else in the workflow so often disappoints people. They open a messy, unstructured Word file, export it with expensive software, and end up with an expensive PDF that is just as untagged as the free version would have been. The tool was never the bottleneck. The habit of building structure first was.
Step 1: Build Structure Before You Export
Open the source file in LibreOffice Writer or Google Docs. Apply actual paragraph styles, Heading 1, Heading 2, Heading 3, instead of just bolding and enlarging text to look like a heading. Add real alt text to every image and chart through the image properties panel, not in a caption underneath. Turn any list that looks like a list into an actual bulleted or numbered list style. Set the document language in the file properties.
Then export. In LibreOffice, use File > Export As > Export as PDF, and in the dialog check the box for "Tagged PDF (add document structure)". In Google Docs, File > Download > PDF Document already produces a tagged PDF using your heading styles, as long as you actually used the built in heading styles rather than manual formatting.
This single habit, structure first, plain export second, solves most of the work that people assume requires paid software.
Step 2: Run PAC 2024 And Read What It Flags
Once you have a PDF, download the free PAC checker and run it against the file. PAC 2024 is a purpose built desktop tool that checks PDFs against the PDF/UA standard and will flag missing tags, broken reading order, untagged images, and table structure problems, all for free, with a plain-language report.
Read the report top to bottom before touching anything. It groups issues by severity, and fixing the "blocking" issues first is usually 80 percent of the work. You can use the checker directly here and rerun it after every round of fixes so you can watch the error count drop in real time.
Step 3: Fix Reading Order And Heading Tags By Hand
For anything the export missed, you'll need a free tag editor. Here's the honest workflow, no shortcuts:
Reading order first. Screen readers follow the tag order, not the visual order. If a two column layout exports with the right column read before the left, that's a reading order bug, not a heading bug, and it has to be fixed in the tag tree directly.
Heading levels second. Check that H1 exists once, H2s nest under it logically, and you never skip a level (H1 straight to H3, for example). A broken heading hierarchy is one of the most common findings in any manual PDF review.
Alt text and table headers third. Confirm every image tag actually carries the alt text you wrote in step one, and that every data table has its header row marked as a table header rather than a plain cell. The W3C's own PDF techniques for WCAG walk through exactly how tagged content should map to these roles, and it's worth keeping that page open while you work through a file the first time.
In our own runs through this workflow, the recurring pattern was that a document with clean heading styles going in needed almost no manual tag editing coming out. The documents that ate up an afternoon were always the ones exported from a source file where headings had been faked with bold text instead of real styles. That's the tell to watch for before you even open a checker.
Step 4: Know When To Still Pay Someone
Free tools close the gap for the vast majority of standard PDFs: reports, brochures, policy documents, most forms. There's a point where paying a specialist still makes sense: complex fillable forms with conditional logic, scanned documents that need OCR plus manual verification, or a backlog of hundreds of legacy files where the hourly cost of manual fixing outweighs a one-time professional cleanup. If your PDF has a table nested inside a table, or a multi-column magazine style layout, budget extra review time or bring in help rather than guessing. Knowing that line in advance saves you from burning a whole afternoon on a file that was never going to yield to a free tool alone, and it keeps you from quietly shipping something half fixed just to hit a deadline.
The same logic that applies to a 40 page PDF applies to something as specific as a scanned restaurant menu, which is its own special case worth knowing about if you ever touch hospitality clients, covered in the guide to restaurant menu accessibility.
A Quick Checklist You Can Reuse
Before any PDF ships, confirm:
- Real heading styles were used in the source document, not bold text.
- Every image has meaningful alt text, and decorative images are marked as artifacts.
- The document language is set.
- Table header rows are tagged as headers.
- The file passes PAC 2024 with no blocking errors.
- Reading order matches the visual layout when tested with tab order or a screen reader pass.
Print that list, tape it above your monitor, and use it every time someone hands you a "quick PDF" on a Friday.
None of this requires a $240 a year subscription. It requires a document built with real structure, a free checker that tells you the truth about what's missing, and about thirty minutes of manual tag cleanup on a normal length document. If you want a hand picking the right free tool for your specific file type, or a second pair of eyes on a document before it goes out the door, our free tools page is a good place to start.
If you'd rather just talk it through with a person, reach our team directly at experts@wcag.world, or head to wcag.world/free-tools to see what's available at no cost right now.