Remediate an EPUB to WCAG 2.2 AA
and get the report that proves it.
Audit against EPUB Accessibility 1.1 and EPUBCheck, repair what can be repaired deterministically, describe the images, and export a conformance report.
PDF remediation — tagged structure, reading order and alt text, in the same pipeline.
Standards audited
WCAG 2.2 AA · Section 508 · EPUB A11Y 1.1 · EAA
Faster than manual
vs. a trained accessibility editor working by hand
AI provider
Azure OpenAI, named in our privacy policy. No one else.
What comes back, and what we leave alone.
A remediated EPUB and the report that accounts for it, line by line. Two of these are findings we deliberately do not fix, because the automatic repair would destroy something.
- Repaired automatically.
Semantics repaired
Rule epub-type-has-matching-roleepub:type and ARIA role reconciled across every document, so assistive technology is told the same thing twice.
- Repaired automatically.
Images described
Rule holoremedi-alt-qualityEach image is read together with the text around it on the page, then described for what it contributes — not for what shapes are in it.
- Reported for your decision.
Credit lines protected
Rule holoremedi-alt-creditAlt text that is an attribution rather than a description is reported, never overwritten. Replacing it would delete the only credit in the book.
- Repaired automatically.
Page list rebuilt
Rule epub-pagelist-brokenPrint pagination restored from the source so a reader can follow a citation to a page number that exists.
- Repaired automatically.
Reading order corrected
Rule epub-toc-orderNavigation re-sorted to spine order, headings returned to document order.
- Repaired automatically.
Front matter resolved without AI
Cover, half-title and title-page images take canonical alt text from the OPF. No inference, so no ambiguity.
- Repaired automatically.
Editorial round-trip
Every image exports to xlsx with its thumbnail. Editors work in Excel; you import the file back and each change is tracked.
- In development.
Tagged PDF structure
Rule pdf-taggingThe same pipeline for PDF — reading order, role mapping, table structure. In development; not part of the service today.
Five surgical phases. One workflow.
Every EPUB moves through the same pipeline. Deterministic where it can be, AI where it adds value, validated at every step.
Triage
EPUBCheck + DAISY ACE + Quality Guard scan every node, image, and metadata field.
Structural Repair
EPUB 2 → 3.3 upgrade, nav.xhtml rebuild, TOC sort, language tag inference.
Front-Matter Fast Path
Cover, title-page, and back-cover images get canonical, deterministic alt text.
AI Alt-Text
Context-aware alt text for every remaining image, generated via Azure OpenAI from the image and the text around it on the page.
Certified Output
Re-validated EPUB + signed audit report exported. Compliance ≥ 99%.
Pay per page and per image.
No subscription. The cost you see before processing is the cost you pay.
Per page
Audit + EPUBCheck + structural repair
Per image
Context-aware AI alt text
Backlist migrations, custom integrations, dedicated support
Contact salesFrequently asked.
We prefer the EPUB 3 page-list nav when publishers provide it. If absent, we count items in the spine. As a last fallback we use word-count / 250 — the print-page convention.
Your first book. On us.
Sign up and get $250 of free credit — enough to remediate one full book end-to-end. No card required, credit never expires.
Start with $250 of credit