PDF OCR — Make Scanned PDFs Searchable

About this tool

PDF OCR converts a scanned PDF — pages stored as flat images with no underlying text — into a searchable, copy-able document by running optical character recognition directly in your browser. Drop your file, optionally click thumbnails to limit which pages are processed (none selected means all), choose a recognition language (English, French, German, Spanish, Portuguese, Italian, or Dutch), and click Run OCR. Tesseract.js renders and reads each page in turn while a progress window shows how far it has got. The recognised text then appears in a panel, headed with a confidence percentage for every page, and pdf-lib adds it to each processed page as one tiny white text block in the top-left corner (up to 8,000 characters per page). The page images are untouched, so the file looks the same, but Ctrl+F, copy-paste and screen readers can find the words — search hits point to that corner rather than each word's spot. Sensitive scans never leave your device.

Why use it

• Searchable Output: The downloaded PDF keeps the original page imagery and adds the recognised text as tiny white type, so Ctrl+F and copy-paste find the words in Acrobat, Preview, and browsers.

• 100% Browser Processing: Tesseract.js and pdf-lib run locally in a worker — your scans never upload to a server, which matters for contracts, IDs, and medical paperwork.

• Multi-Language Recognition: Switch between English, French, German, Spanish, Portuguese, Italian, and Dutch with one dropdown — the right language model dramatically improves accuracy.

• Page Picker: Process only the pages that matter — handy for 200-page document dumps where you only need text from chapters 3 and 7.

• Per-Page Confidence: The text panel labels every page with Tesseract's confidence score so you can spot pages that need a sharper rescan.

• Free with No Watermark: Convert as many scans as your laptop can chew through — no daily quota, no signup, no branded output.

How to use

1. Drop or select a scanned PDF — the tool reads its page count and shows a thumbnail picker.
2. Click thumbnails (or use All / Odd / Even / Invert) to limit which pages are processed — leave none selected to OCR every page.
3. Choose the document's primary language so Tesseract loads the matching trained data file.
4. Click Run OCR — a progress window shows which page is being rendered and recognised.
5. Review the recognised text in the panel below, where each page is headed with its confidence percentage, and copy it if needed.
6. Click Download Searchable PDF to save a copy with the recognised text added to each processed page.

When it helps

• When a scanned contract or PDF arrives by email and you need to search for clauses or names without retyping.
• When archiving paper receipts and tax slips into a searchable PDF library before tossing the originals.
• When making old academic books or articles accessible to screen readers and full-text search.
• When preparing legal discovery materials so attorneys can grep for terms across thousands of scanned pages.
• When digitising handwritten notes or printed meeting minutes so they can be indexed by note-taking apps.
• When you receive a scanned ID or passport copy that needs to be parsed without uploading the file to a third-party service.

Examples

Tips

• Higher-resolution scans (≥ 300 DPI) produce dramatically better OCR results — rescan low-quality faxes if you can.
• If a page mixes two languages, pick the one with more text — Tesseract handles single-language documents most accurately.
• Crop or deskew the original before running OCR; even 2–3° of skew tanks recognition accuracy.
• Process big PDFs in batches of 20–30 pages at a time on mobile devices to avoid running out of memory.
• After downloading, verify by searching for a phrase you know is on page 1 — if it doesn't match, the language or scan quality probably needs adjusting.

Frequently Asked Questions

Glossary

OCR

Optical Character Recognition — the process of converting page images into machine-readable text.

Searchable PDF

A PDF that retains its original visual layout but also embeds a text layer so the contents can be searched, copied, and read aloud.

Text Layer

Machine-readable text stored in a PDF alongside its images; here it is one small block of white text per page rather than word-by-word positions.

Tesseract.js

A JavaScript port of Google's Tesseract OCR engine that runs entirely in the browser via WebAssembly.

Confidence Score

A percentage Tesseract assigns to each recognised page indicating how certain it is of the read.

DPI

Dots per inch — the resolution of a scanned image; ≥ 300 DPI is the practical floor for reliable OCR.