How to Make a Scanned PDF Searchable (and Get the Text Out) With OCR

Is your PDF text or a picture?

Try to select a single word. If it highlights on its own, the PDF already contains real text — you don’t need OCR; use PDF to text to copy it all out, or PDF to Word to edit it. If the whole page highlights as one block, or nothing highlights, it’s an image, and OCR is the answer.

How to extract text with OCR

1. Open PDF OCR and choose a scanned PDF or a photo (JPG, PNG, WebP or iPhone HEIC).
2. Choose the document language — 46 are available, from English, Arabic and Chinese to Hindi, Vietnamese and Swahili.
3. For a PDF, optionally enter the pages to read (up to 50 per run).
4. Start recognition. The first run downloads the OCR engine and language data, which takes a moment; later runs are faster.
5. Download the searchable PDF, and/or check the recognised text in the box, then copy it or download it as a .txt file.

What you get

• A searchable PDF. Your original pages, untouched, with the recognised text added invisibly behind each one. Ctrl+F finds words, you can select and copy sentences, and file search on your computer can find the document by its contents. It adds only a few kilobytes per page.
• The text itself, to paste into an email, a form, a spreadsheet or a Word document, or to download as a .txt file.

OCR doesn’t rebuild the page’s layout, fonts or tables. For a document you need to edit and reuse, the usual workflow is: copy the recognised text into Word or Google Docs, then add headings and formatting.

Getting accurate results

Proofread the result

Even with a perfect scan, OCR makes small mistakes, and they tend to be the dangerous kind: numbers and names. Look especially for:

• Digits confused with letters: 0 and O, 1 and l and I, 5 and S, 8 and B.
• Decimal points and commas dropped from amounts.
• Line breaks in the middle of sentences, where the original had a line end.
• Hyphenated words split across lines (“docu- ment”).
• Columns read across the page instead of down — common with two-column layouts.

For anything important — an amount, an account number, a date — compare the OCR text with the original image before using it.

Uses that save real time

• Quoting a clause from a scanned contract without retyping it.
• Filling in a form with details from a scanned letter.
• Turning a printed handout into an editable document you can update.
• Copying receipt amounts into a spreadsheet for an expense claim.
• Reading text in another language — OCR the text, then paste it into a translation service of your choice.
• Extracting text from screenshots, such as an error message or a slide.

Multi-page documents

When you OCR a PDF with several pages, the result is divided with markers such as “--- Page 3 ---”, so you can find where each page’s text begins. For long documents, run the pages you need rather than the whole file — it’s quicker, and each run handles up to 50 pages. If a document mixes languages page by page (an English letter with a Spanish attachment, say), run each part with its own language setting.

Tables and forms

OCR reads text line by line, so a table comes out as rows of words separated by spaces rather than neat cells. For a small table, paste the text into a spreadsheet and use Data → Text to Columns (in Excel) or Split text to columns (in Google Sheets) with spaces as the separator, then tidy up. For forms, it’s often quicker to copy only the values you need than to reconstruct the whole layout.

OCR on a phone

The tool works in mobile browsers, and phone photos are a common source: a sign, a whiteboard, a page in a book. Take the photo in good light, straight on, filling the frame with the text, then open it in PDF OCR — HEIC photos from an iPhone work directly. Many phones also have text recognition built into the camera or photo app, which is handy for a quick line or two; for whole pages and documents, a dedicated OCR run with the right language set is more reliable.

How a searchable PDF works

The recognised words are placed as invisible text exactly over the words in the page image. Viewers show you the image, but when you search or drag to select, they use the invisible text — which is why selection highlights line up with the words you see. Because the text comes from OCR, a misread word in the text layer won’t be found by searching for the correct spelling, so for critical documents it’s still worth proofreading the extracted text.

Saving what you extract

Download the text as a .txt file named after the original — “Lease-2026-OCR.txt” next to “Lease-2026.pdf” — so you can find both later. Plain text opens on every device and pastes cleanly into anything.

Worked example: expense receipts

Hana has 14 photographed receipts from a work trip and needs the dates, suppliers and amounts in her company’s expense spreadsheet.

1. She combines the photos into one PDF with JPG to PDF, so she has a single file to attach to the claim.
2. She runs that PDF through PDF OCR with English selected — all 14 pages in one go.
3. From the recognised text she copies each date, supplier and total into the spreadsheet.
4. Two amounts look odd — “$4B.50” and “$1O2.00”. Checking the photos, they’re $48.50 and $102.00. She corrects them.

Typing everything by hand would have taken about half an hour; OCR plus checking took ten minutes, and the check caught the two errors that mattered.

Privacy

Recognition runs in your browser: the engine and language files are downloaded to your device, and your document is processed there. Receipts, contracts and letters aren’t uploaded.