Make every PDF readable. By everyone.
AI-assisted accessibility tagging for PDFs. doclarity does the heavy lifting; you review and refine — verified against PDF/UA and WCAG, plus an AI check for what other tools miss.
The European Accessibility Act applies since 28 June 2025. Public-facing documents have to work for every reader. doclarity helps you bring PDFs up to standard — and shows its work.
Three steps, and you stay in control.
Upload a document
Word, PowerPoint, LibreOffice, or PDF — doc, docx, ppt, pptx, odt, odp, pdf — born-digital or scanned. OCR mode and automatic page rotation bring the paper era along. Office files are converted on the way in; the result is always a PDF.
doclarity rebuilds the structure
It strips broken tags, builds a fresh structure tree, drafts alt text, corrects reading order and heading levels, and fixes text contrast to your WCAG target.
Review, verify, download
Fine-tune anything in the review editor. Then download the tagged PDF together with its veraPDF validation report.
Structure, not guesswork.
Existing tags are often worse than none. doclarity starts clean: it wraps the original content, rebuilds the structure tree from what the document actually says, and keeps heading levels consistent from the first page to the last.
- Headings, paragraphs, lists, tables, footnotes, links — thirteen element types
- Recurring logos and watermarks are announced once, not on every page
- Fonts and bookmarks repaired along the way
You have the final word.
AI drafts; you decide. The review editor puts every element next to the real page: change a type, rewrite alt text, drag the reading order, set table header cells, insert what the AI missed. Undo included.
- Side-by-side: element list and PDF with color-coded regions
- Table, list, and link editors for the fiddly parts
- Keyboard shortcuts for reviewers who mean it
Contrast, fixed and documented.
Pick your target — A, AA, or AAA. doclarity measures every text run against its actual background, moves colors the smallest distance that passes, and records each change as a before / after with the ratios to back it up.
- Region-aware: white text on a dark box stays white
- Large-text thresholds applied where they belong
- Anything unfixable lands on a manual-review list, not under the rug
Thorough where it counts.
Alt text that says what matters
AI drafts a description for every figure — what it shows, not what it is. You approve, edit, or rewrite. Nothing ships unseen.
Reading order that follows the story
Two columns, sidebars, footnotes: the reading path follows meaning, not coordinates on the page.
Tables, forms, links
Header cells with proper scopes, labeled form fields, links that lead where they say. The fiddly parts, done right.
Scanned paper, welcome
OCR mode reads what the scanner saw; automatic rotation straightens what the feeder tilted. The archive comes along.
Verified, not assumed
Checked three ways — veraPDF for PDF/UA, WCAG success criteria, and an AI review of the semantics checkers miss. The proof is part of your report.
Built for teams
Organizations with shared documents and roles. German and English, light and dark, credits that are easy to reason about.
Proof, not promises.
Every tagged document goes through three layers of checking. veraPDF validates the machine-readable PDF/UA structure (ISO 14289). A WCAG pass covers the success criteria a document has to meet. And an AI review reads the document the way a person would — catching the semantic problems traditional checkers simply can't see. One report, ready for auditors, procurement, or your future self.
- veraPDF — PDF/UA-1 and PDF/UA-2 (ISO 14289)
- WCAG 2.1 success criteria, checked and reported
- AI semantics: is the alt text meaningful, is the reading order sensible, are headings a real hierarchy
Ready for clarity?
doclarity is in its test phase — access is granted personally, and your feedback shapes the tool.