Skip to main content

What it does & how secure it is

Every capability of the browser tools, and the privacy model behind them.

Home — all tools

What the tools do — right in your browser

Fixes automatically

  • Document title, language, and metadata
  • Heading order and bookmarks
  • Table header cells and scope
  • Link descriptions and link tags
  • Form field tooltips
  • Reading and tab order settings
  • Repeating headers/footers hidden from screen readers
  • Low-contrast text darkened (optional)
  • Office files: titles, language, table header rows, spacing paragraphs

Adds missing structure

  • Auto-tags untagged PDFs — headings, paragraphs, lists, tables, and figures
  • OCR gives scanned pages real, selectable text (English runs fully on-device)
  • Optional on-device AI drafts alt text for your images — nothing is uploaded
  • Marks Excel data ranges as real tables with header rows

Guides you through the rest

  • One-click “Fix everything automatically” runs OCR, tagging, and AI alt-text drafts in the right order on each file
  • Step-by-step fix instructions using only free tools
  • Whole-folder drops — every PDF or Office file inside is processed
  • “Show on page” highlights exactly where each problem is
  • Fill-in panels for alt text, headings, table headers, and links
  • Reading-order editors for PDFs and PowerPoint slides
  • Tag structure viewer shows the document the way assistive technology sees it
  • Before-and-after scorecard and a downloadable remediation record on every file
  • Listen mode reads your document the way a screen reader would

Everything above runs in this browser tab — your documents are never uploaded, and the browser build requires an internet connection to load its tools. Need to convert Word or PowerPoint files to tagged PDFs, or batch hundreds of documents? Get the free desktop app.

Temporary work only

This browser build has no save-and-resume or offline installation. Use Clear files to release the workspace, and download copies explicitly before leaving. Earlier app caches are retired where browser permissions allow.

How secure is this?

Your documents never leave your device

The browser tools do every step — reading, checking, fixing, OCR, and the optional AI descriptions — with code running inside this tab. No document, page image, or extracted text is sent to a server by the document tools. The app keeps documents and review decisions in temporary page memory, and does not save them to browser storage. Clear files, leave the workspace, or close the tab to release that state. Explicit downloads remain wherever you save them. This is not a guarantee of secure erasure from browser internals, extensions, or operating-system memory.

What the site does download (never upload)

  • The app itself, its fonts, and the OCR engine — all served from this site.
  • If you turn on AI descriptions: the AI model (~220 MB for each session) from the Hugging Face open-source library. The model comes down to you; your images never go up.
  • OCR in Spanish, French, or German fetches that language pack from a public library when needed. English is served by this site.
  • Model and OCR data are not kept in an app-managed persistent cache.

No account, no tracking by the tool

No login, no signup, no keys to manage. The document tools do not send document data to analytics or error reporting. The site is hosted on the Lovable platform, whose standard page hosting applies, including ordinary access logs for page and asset requests — but your documents are processed only by the code in this tab and are never part of any request.

Open source and auditable

Both versions are built entirely from widely used open-source components (pdf-lib, Mozilla’s pdf.js, Tesseract OCR, and more) — nothing proprietary or hidden. The full component list, licenses, and security details are published in the Security & Privacy overview on GitHub.