Free accessibility tools for public institutions
Make every document accessible.
Browser-based tools to test, convert, and validate documents against WCAG 2.1 and PDF/UA. Built for universities and government teams. Nothing to install.
Checks against WCAG 2.1 and PDF/UA. To pilot or report a problem, email deepa@equitabledocs.org.
01 Tool one
QCDoc.
PDF accessibility audit made simple.
In testing
A free accessibility assessment for the documents your students rely on. Upload a PDF; receive a report in three parts: a plain-language summary, a technical report mapped to WCAG and PDF/UA, and a compliance dossier built to pass independent validation. Last-mile remediation by EquitableDocs if the document needs human work.
What you get:
- Reports on all 31 PDF/UA-1 checkpoints (Matterhorn Protocol) and WCAG 2.2 AA: it machine-checks every condition a machine can verify, flags the rest for human review, and scores the two standards separately.
- A walkthrough that teaches the standards as it goes, and a standards atlas recording every checkpoint with a captured image of each finding, citable for procurement.
- A remediation pack for the team doing the fixes (an annotated PDF, a manifest, and a machine-readable JSON file).
- A printable compliance report you can file or share.
- A logical-structure inspector showing the document's tag tree, and a screen-reader preview that labels what a screen reader announces page by page.
- Routing for documents that need human work (STEM equations, complex diagrams, scanned PDFs without a text layer) to the EquitableDocs remediation workflow.
- Built so non-specialists can act on the report. Specialists already have veraPDF and PAC; this tool is the bridge for everyone else.
02 Tool two
AccessMitra.
In testing
A free public PDF remediation tool. Upload an inaccessible PDF; AccessMitra runs it through the full pipeline (font fixes, tag structure, alt text for figures, table headers, math handling, language metadata, validation) and emails back an accessible tagged PDF, an accessible Word file, and an EPUB.
How it differs from QCDoc. QCDoc reads a PDF and tells you what is wrong with it, in plain language a non-specialist can act on. AccessMitra goes a step further: it fixes what it can fix automatically and returns ready-to-share accessible files. Run a PDF through it first; if the issues are remediation-shaped, run the same file through AccessMitra.
What you get:
- Anonymous upload, no account, file deleted two days after upload.
- Auto-tagging via PDFix SDK with document-family-specific presets where available.
- AI-generated alt text for every figure, with confidence scores; below threshold the file is escalated to human review.
- Equation handling via Mathpix where present, with MathML output and plain-language alt text.
- Coverage report against the Matterhorn Protocol, PDF/UA-1, and WCAG 2.1 AA, plus a manifest of every change made.
- Two paths: simple text-based PDFs deliver straight from the automated output; documents with STEM, complex diagrams, or scanned images go through human verification before delivery.
- Outputs delivered: tagged PDF, Word, and EPUB, by email link.
03 Tool three
AltBridge.
In testing
A browser tool that helps remediators and subject specialists agree on alt text for images. The remediator uploads a tagged PDF; the tool extracts each figure, drafts an AI suggestion, and creates a shareable link. The specialist opens the link, confirms or rewrites each suggestion, and adds notes where guidance is needed. Responses export as one CSV the remediator loads back into their workflow.
What you get:
- Every figure detected in the source PDF, including its current alt text when present, with a magenta highlight on the rendered page so the reviewer sees the figure in context.
- A plain-language alt text draft from Claude, with a verdict (match, ambiguous, mismatch), confidence, and reasoning.
- Approve, rewrite, or skip each figure. Responses save in the browser as you work.
- Exports as a CSV; also produces an accessible Word file with the reviewed alt text embedded on each figure.
- Works end-to-end with a screen reader. Visible labels, keyboard operation, colour-supported announcements.
04 Tool four
Font Auditor.
In testing
A free font check for any PDF. Some fonts cause text to disappear, garble, or fail screen reader software. The Font Auditor reads your file, lists every font, and tells you which ones are safe and which to replace before sharing or compressing. It also checks for legacy non-Unicode fonts in Indic scripts (a common cause of unreadable text in Hindi, Tamil, Kannada and other regional-language PDFs). Built for designers, typesetters, faculty, library staff, and anyone preparing a PDF for distribution.
What you get:
- Every font on every page and every embedded form classified as Fail, High impact, Low impact, or Pass, in two sections (Text and language readiness; Rendering and embedding).
- Standard 14 PostScript names (Helvetica, Times-Roman, Courier, Symbol, ZapfDingbats) flagged when referenced without an embedded font program.
- The case where an embedded subset is too small to cover the document's character set caught early; this is the silent failure that shows up as missing letters or blank squares.
- Specific catches for legacy non-Unicode fonts in Indic scripts (Walkman-Chanakya for Devanagari, Nudi for Kannada, Aryan2 for Hindi, and others). The tool also catches the harder case where the font is not in our catalog but the extracted text is gibberish. For the full picture of why this matters and how to fix it at the source, see Lipi.
- A specific replacement named for every problem found: Arial or Source Sans 3 for Helvetica; Times New Roman or Source Serif 4 for Times-Roman; Courier New or JetBrains Mono for Courier.
- A downloadable Markdown report and a JSON summary.
05 Tool five
Accessibility-Safe Compressor.
In testing
A free PDF compressor that makes files smaller without breaking screen reader access. Most compressors strip embedded fonts, drop tags, or flatten the structure tree; the result looks the same in a browser but a screen reader cannot read it. This tool compresses by recompressing images and re-packing the PDF object stream, never touching the structure tree, fonts, alt text, RoleMap, language tags, or bookmarks.
What you get:
- A font audit first, using the same logic as the Font Auditor. If any font is BROKEN, the tool refuses and tells you what to fix at the source. Compression cannot fix a font problem; the source document has to.
- Images decoded, oversized images resized, and re-encoded as JPEG at quality 80. Images with soft masks or stencil masks are skipped to preserve their alpha channels.
- The PDF object stream re-packed and Flate streams recompressed at maximum level.
- Every accessibility feature preserved byte for byte. Verified on the canonical CABE newsletter sample: 1.72 MB compressed to 1.50 MB (12.6 percent saved), 106 of 106 PDF/UA-1 rules still pass after compression.
06 Tool six
Page Accessibility Coach.
Developer alpha (v0.1)
A Chrome extension that runs an accessibility scan on any web page and translates the findings into plain language using the Claude API. No WCAG jargon, no node selectors, no engineering vocabulary. Built for faculty, library staff, and anyone who creates or maintains web content.
How it differs from QCDoc. QCDoc reads PDF, Word, and PowerPoint files. The Coach reads web pages, in your browser, in real time. The translation step (Claude rewriting WCAG findings into plain English) is the same in spirit.
What you get:
- One-click scan of the active tab, using axe-core (the same engine behind axe DevTools).
- Each finding translated by the Claude API into a one-sentence "what is wrong" and a one-paragraph "how to fix".
- Severity badges: Blocker, Hindrance, Polish.
- Copy report as Markdown for pasting into email or a pull request.
- Runs locally. Your Anthropic API key stays on your machine; the page URL and findings only go to api.anthropic.com.
v0.1 is a developer-install alpha. Install steps are in the repository README.
07 Companion knowledge base
Lipi.
The language-readiness foundation for non-English content.
Live
A knowledge base on making documents accessible in the languages of the Global South. Lipi explains the gap, legacy non-Unicode fonts (Walkman-Chanakya, Nudi, Aryan2 and the rest), the four conditions a regional-language file has to meet, and how to remediate a regional-language book end to end. It is the open companion to the Font Auditor: the Font Auditor catches what is catchable from font names and gibberish signals; Lipi explains the rest, including the cases that need a human eye.
What you find on Lipi:
- Why regional-language PDFs in Devanagari, Tamil, Kannada, Bengali, Urdu and other Indic and Asian scripts fail when they look right on screen.
- What legacy non-Unicode fonts are, how to spot one, and how to fix the document at the source.
- The four conditions a non-English document has to meet for a screen reader to read it.
- A step-by-step for remediating a regional-language book.
07 Feedback and pilots
Pilots inform the roadmap.
These tools are open to pilot use by remediation teams, university disability-services offices, library teams, and publishers, in the Global South and elsewhere. To discuss a pilot, request a real session for your batch, or raise a bug, email deepa@equitabledocs.org.