OCR PDF — Make a Scan Searchable
Add a real, selectable text layer to a scanned PDF or photos of documents — free and private. The page looks the same, but you can finally select, copy and search the words. You also get a plain-text file.
Recognition runs in your browser — your document is never uploaded.
Click to choose a scanned PDF or drag and drop it here
Or select several images (JPG, PNG) to turn into one searchable PDF
How it works
Each page is first rendered to an image with pdf.js, the same engine Firefox uses to display PDFs. That image is passed to Tesseract.js — the WebAssembly build of Google's Tesseract OCR engine — which returns the recognised words together with the pixel box each one occupies. The tool then rebuilds the PDF with pdf-lib: it draws your original page image untouched, and over each word it places invisible text at the matching position. Because the text is rendered in invisible mode it never alters how the page looks, yet it can be selected, copied, searched and indexed. To keep that hidden text faithful when you copy it, a purpose-built glyphless font is embedded, which avoids the garbled-copy problem that plagues naïve text layers.
How to make a scanned PDF searchable
- Click the box above or drag a scanned PDF (or images) onto it.
- Choose the language of the document.
- Click Make searchable and watch the per-page progress.
- Save the searchable
PDFand the plain-text.txtfile.
An honest note on speed and accuracy
Running OCR locally is what keeps your document private, but it is the trade-off for speed: the work happens on a single thread of your own device, so a long, image-heavy scan can take a while — the progress bar is per page so you always know where it is up to. Accuracy is honest, too. A crisp, well-lit scan in the correct language reads remarkably well; a faint fax, a skewed photo or handwriting will read poorly. Choosing the right language helps a lot, and the bundled text file lets you eyeball the result and fix anything important before you rely on it.
Why OCR in your browser instead of uploading
Searchable-PDF services almost always upload your file to a server to run recognition, which means handing over the very documents — contracts, statements, records — you most want to keep to yourself. This tool never does that: the recognition engine runs on your machine and your file stays on your machine. It is free, requires no account, adds no watermark, and after the first visit the page and engine are cached so it keeps working offline.
Related tools
New here? Read how to merge pdfs without losing quality or browse all our PDF guides.
Frequently asked questions
What does OCR do to my PDF?
It reads the text in your scanned pages with optical character recognition and adds an invisible, selectable text layer on top of the original image. The page looks identical, but you can now select, copy, search and index the words.
Is my file uploaded to a server?
No. The pages are rendered with pdf.js and recognised with Tesseract.js entirely in your browser. The recognition engine and language data are fetched from a public CDN, but your document itself is never uploaded or stored anywhere.
Why is it slow on long documents?
To keep your file private, recognition runs locally on a single thread of your device rather than on a server farm. That is reliable but not fast, so a long scan can take a while. A per-page progress bar shows where it is up to.
Which languages are supported?
You can choose English, French, German, Spanish, Italian or Portuguese. The matching language data is downloaded the first time you use it and then cached. Picking the right language noticeably improves accuracy.
How accurate is it, and what do I get back?
Accuracy depends on the scan: clean, high-contrast pages read very well, while faint, skewed or handwritten pages read poorly. You get a searchable PDF plus a plain-text .txt of everything that was recognised, so you can check or reuse the text.