Make a scanned PDF searchable.

Reads the words off the page and writes them back as an invisible text layer, so the document looks identical but Ctrl-F finds things. It all happens on this machine. The recogniser and its English language data are staged inside the app and served from this site — about 6 MB on your first run, from us and never from anyone else’s CDN. Your document is the thing that goes nowhere at all.

$99 once and the whole toolkit is yours — Acrobat is about $155 a year and stops working the day you stop paying. And the box of scanned letters you are working through stays on your laptop: we never see them, because nothing you open here is stored on a server.

Questions people actually ask

Where does the file go?: Nowhere. The recogniser, the English language data and the PDF writer all run inside this page, on your machine, so we never see it. The app is served with a Content-Security-Policy that only permits connections back to its own origin, which means the browser refuses to let this page open a connection anywhere else — you can watch the network tab while it works.
So nothing is downloaded either?: That part is not true, and we would rather say so than let you find out in the network tab. The first run fetches the recogniser — a WebAssembly core and the English training data, about 6 MB together. What matters is where from: PDFMod's own origin, staged into the build, never a third-party CDN. Tesseract's default is to pull those files off jsDelivr, which would tell a stranger the moment you were about to OCR something, so all three paths are pinned in code and connect-src 'self' would refuse anything else anyway. After that first run they are cached and OCR works with the network off. Your document is the thing that never moves.
What does “invisible text layer” mean?: Your scan is a picture of words. OCR reads that picture and writes the words back onto the same page in a font that draws nothing. The page looks pixel-for-pixel the same, but Ctrl-F finds things, you can select and copy, and a screen reader can read it.
Will it change how my document looks?: No. Nothing is re-drawn, re-compressed or rebuilt — the original page objects are kept exactly as they are, and the recognised words are drawn over them in a font that marks no pixels at all. Nothing of yours is covered up, because nothing of ours is visible. Form fields, links and bookmarks survive.
How accurate is it?: On clean printed pages, typically 93–95% mean confidence, and it holds up on genuinely poor scans — we tested skew, blur, speckle and heavy JPEG artefacts. It is still a reading machine: handwriting, unusual fonts and very faint print will contain mistakes. You can read everything it recognised before you download.
Which languages?: English today. Every extra language is another 3MB shipped inside the app, so they get added deliberately rather than fetched on demand from someone else's CDN.

OCR adds a text layer to a scan; it does not let you rewrite the page. Changing words a document already carries is a different tool — Edit — and it is in this toolkit, with its own limits stated on its own page.