PDF remediation for every reader. No exceptions.

Remediate turns the PDFs you already publish into screen-reader-ready documents — with structured headings and tables, OCR for scans, translation and cited answers. Meet WCAG 2.1 AA for the ADA Title II deadlines without re-authoring the archive.

Start free trial Open the live demo

One script tag No re-authoring WCAG 2.1 AA & PDF/UA structure

lonestarcountytx.gov/auditor Remediate
The Remediate viewer showing the Lone Star County annual financial report as an accessible transcript, with a document outline, translation and question panel in the sidebar.
The 2024 Lone Star County ACFR — 39 pages, read as a structured transcript. Open this document

Designed against the standards your compliance office already cites

  • ADA Title II
  • WCAG 2.1 AA
  • Section 508
  • WAI-ARIA

01 Why it matters

A PDF can look perfect and still be unreadable.

Screen readers do not see the page. They read the structure inside it. When that structure is missing, a 39-page financial report collapses into a stream of fragments — no headings, no table relationships, no idea what the figures mean.

A

Tagging is not the same as reading.

A PDF can be “tagged” and still announce every number without its column, or every figure without its caption. Remediate maps the document into a transcript that preserves the relationships a reader needs.

B

The deadline is not abstract.

Under the DOJ’s ADA Title II rule, public entities must meet WCAG 2.1 AA for web content, including the documents published on their sites. Entities serving 50,000 people or more have until April 26, 2027; smaller entities and special districts have until April 26, 2028.

C

Re-authoring every file is not a plan.

Most accessibility projects stall because remediating an archive by hand takes years. Remediate works with the files already on your site, and leaves the original PDF untouched for anyone who needs it.

Screen reader output
graphicimage, 612 × 792
textCounty 2024 □□ □□ Table 1 2 3 4
linkhere
table4 columns, 12 rows — no headers announced
graphicimage

Reading order lost. Context gone. The page is technically text, but practically silence.

Illustrative screen-reader output for the same page.

1 in 4

U.S. adults lives with a disability. For many, an untagged PDF is a closed door.

CDC Disability and Health Data System

Apr 26, 2027

Compliance date for public entities serving 50,000 people or more under the ADA Title II rule, extended a year by DOJ in April 2026.

DOJ interim final rule, 28 CFR Part 35

10 languages

Spanish, Chinese, Vietnamese, Arabic, French, Korean, Russian, German, Tagalog and Haitian Creole — in place.

Built into the viewer

02 How it works

From published PDF to reading experience, in four steps.

Remediate sits between the files you publish and the people who need them. Nothing about your publishing workflow has to change.

  1. 01

    Publish as usual

    Keep your CMS, your file names and your links. Remediate recognises the PDFs on the page and works from the originals.

  2. 02

    Build the structure

    Headings, reading order, tables, figures, form fields and footnotes become a semantic transcript that mirrors the source.

  3. 03

    Review the exceptions

    Alt text, table headers, language and callouts are visible in one place, so a reviewer can fix what automation cannot.

  4. 04

    Embed the reader

    One script tag upgrades every PDF link on the page. The modal, deep links, keyboard support and screen reader announcements are handled.

03 What's inside

A reading experience, not a viewer.

Everything a visitor needs to understand a document — without downloading a plugin, printing a page, or asking someone else to read it to them.

Structure

Structure you can navigate

The transcript mirrors the report: headings, lists, tables, figures and page landmarks. Screen readers land where they expect, and the outline is real — not a list of guesses.

  • Real heading levels and landmarks
  • Tables with captions and header cells
  • Alt text and figure captions
A financial table inside the Remediate transcript, with labelled columns, a caption and readable figures.
Translation

Translation that keeps the context

The whole transcript translates in place — including tables and headings. The original PDF stays in its source language, and a note says so.

  • Español
  • 中文
  • Français
  • العربية
  • Tiếng Việt
  • Kreyòl
  • 한국어
  • Русский
  • Deutsch
  • Tagalog

Machine translation, reviewed by exception. Direction and language attributes are set for every passage.

Reading aids

Reading options for the way people actually read

Lexend for dyslexia-friendly reading, extra line spacing and high-contrast inversion. Applied to the transcript in one click.

Answers

Answers with receipts

Ask a question in plain language. Remediate answers from the document and shows the page and element it came from, so nobody has to take a machine's word for it.

Try it in the demo
Find & display

Find anything, go anywhere

Search the full transcript, jump to any page and control the layout without leaving the keyboard.

Every screen

The same document on a phone

A two-row toolbar, readable type and a transcript that reflows — not a PDF shrunk to fit.

The Remediate transcript on a phone, showing the accessible document title and reading options.

04 Standards

Built to the standard your compliance office cites.

Remediate targets WCAG 2.1 AA and the ADA Title II document rule. The original PDF stays available, and every accessible feature is keyboard operable.

Perceivable

A semantic transcript, alt text for figures and captions for tables. Nothing relies on colour alone.

WCAG 1.1, 1.3, 1.4

Operable

Every control is reachable by keyboard, with skip links, visible focus and document shortcuts.

WCAG 2.1, 2.4

Understandable

Consistent navigation, plain labels, page landmarks and an outline that matches the document.

WCAG 3.1, 3.2

Robust

Landmarks, heading levels and live announcements that hold up across screen readers and browsers.

WCAG 4.1

The ADA Title II clock

  1. Apr 24, 2024 DOJ publishes the web and mobile app accessibility rule for public entities.
  2. Apr 20, 2026 DOJ publishes an interim final rule extending both compliance dates by one year, effective immediately.
  3. Apr 26, 2027 Compliance required for public entities with a total population of 50,000 or more.
  4. Apr 26, 2028 Compliance required for smaller public entities and special district governments.

Dates from the DOJ final rule (28 CFR Part 35), as extended by the interim final rule published April 20, 2026. Documents published on a public website are part of that web content. Consult your counsel for your entity’s obligations.

05 Live demo

Don't take a screenshot's word for it.

The demo is the actual product, running on four documents from a mock County Auditor site. Open a report, switch to the transcript, translate it, ask it a question. No signup, no video.

Browse the demos Open the county site
Live demo
lonestarcountytx.gov/auditor Remediate
The mock Lone Star County Auditor website with the Remediate document viewer open over it.
The viewer opens over the page that published the PDF — same link, same file, accessible by default.

Synthetic financial data. Real viewer behaviour.

06 Integration

One script tag between you and an accessible archive.

Remediate works with the links you already have. Add the script to your template and every PDF link on the page becomes an accessible reader — modal, deep link, keyboard and screen reader included.

  • No changes to your document URLs
  • Deep links survive the modal
  • Works with any CMS or static site

Document processing runs as a service. In this build, npm start serves the whole demo locally.

<!-- Add once, in your site template -->
<script src="/docbox.js" async></script>

<!-- Your PDF links stay exactly as they are -->
<a href="/docs/2024-acfr.pdf">2024 Annual Report</a>

07 Guides

PDF remediation guides for ADA and Section 508 teams

The deadlines, the standards and the review steps — written for the people who own public documents, not for search engines.

Pillar guide PDF remediation, without re-authoring the archive What remediation is, which standards apply, how the workflow runs and what drives cost. Read the guide Deadline guide ADA Title II and the 2027–2028 PDF deadlines Which entities are covered, what counts as web content, the exceptions and what to do now. Read the guide Federal guide Section 508 PDF remediation for federal teams 36 CFR 1194, WCAG 2.0 AA, VPATs, testing steps and where remediation fits. Read the guide Working checklist The PDF remediation checklist Structure, tables, forms, scans, metadata and verification — copy-ready for reviewers and vendors. Open the checklist

All PDF accessibility resources

08 Questions

The things everyone asks first.

If your question is not here, the demo is the fastest way to answer it.

See it for yourself Four documents, every feature
What is PDF remediation?

PDF remediation is the process of making an existing PDF readable by screen readers and other assistive technology: repairing tags and reading order, adding headings, table headers, alt text, form labels, language metadata and a text layer for scans. It changes the structure behind the page, not the visual design.

When is the ADA Title II deadline for PDFs?

Public entities with a total population of 50,000 or more have an ADA Title II compliance date of April 26, 2027. Smaller public entities and special district governments have until April 26, 2028. DOJ extended both dates by one year in April 2026. The technical standard is WCAG 2.1 Level AA. Read the deadline guide.

How much does PDF remediation cost?

Cost depends on whether documents are scanned, how many tables and forms they contain, the review required and the volume. One plan covers every document, page and seat, quoted with your team after a short conversation about your publishing programme. Talk to sales or read about cost drivers.

Does Remediate replace our PDFs?

No. The original file remains the source of truth and stays downloadable. Remediate adds an accessible transcript and a reader alongside it, so the document looks and behaves the way people expect while becoming usable by assistive technology.

Do we have to re-author every document?

No. The transcript is generated from the published file and mapped to its structure. Reviewers spend their time on the exceptions automation cannot resolve — an ambiguous table header, a decorative figure, a scanned page — rather than rebuilding the archive by hand.

What about scanned documents?

Scanned pages need text recognition before they can be read. The demo includes a document in that “transcript on request” state, so you can see how a county can ask for a transcript and have the accessible version appear without changing the published file.

Is the viewer itself accessible?

That is the point. The viewer supports keyboard-only operation, visible focus, skip links, landmarks, live announcements and announced state changes. Its automated test suite checks those behaviours on every build, including hit-testing, scrolling and focus restoration.

Which languages can it translate?

Ten today: Spanish, Simplified Chinese, Tagalog, Vietnamese, Arabic, French, Korean, Russian, German and Haitian Creole. Translation is machine-generated and labelled as such, with proper lang and dir attributes; the original language always remains one click away.

How is it deployed?

The browser client is a single static script. Document processing runs as a service that produces the transcript, thumbnails and metadata. This build bundles both: npm start seeds the demo documents and serves the viewer, the site and the API on one port.

Is the demo data real?

No. The Lone Star County site is a mock, and its financial figures are synthetic sample data, clearly labelled inside each document. The document structure, PDF rendering, transcript, translation and viewer behaviour are real.

09 Start here

The next document you publish can be readable by everyone.

Start with one report. See exactly what your readers would get.

Start free trial Open the live demo