Can't copy text from a PDF? Why it happens and how to get the text

If you cannot select words in a PDF, the pages are most likely pictures of text, a scan or a photo, rather than text. Drop the file into PDF to text: if it finds words, copy them from there. If it reports that no text was found, the pages need OCR (optical character recognition), which reads the letters from the picture. Both steps run in your browser, and the file is never uploaded.

PDF to textNeed the text out of a PDFDo it free

Text PDF or scanned PDF

A text PDF stores each letter as a character, with a font and a position on the page. A scanned PDF stores each page as one image, the way a photo does, so there is nothing to select. Some files mix the two: typed pages with a scanned signature, or a scanned form with a typed title.

Test it in your viewer by trying to select one word. If the highlight follows the letters, the text is there. If you can only select the whole page, or nothing at all, the page is an image.

What you see	What it usually means	What to do
Words select and copy	A text PDF	Copy directly, or use PDF to text for the whole file
Nothing selects, and PDF to text finds no text	Scanned or photographed pages	Render the pages, then run OCR
Words select but paste as odd symbols	The fonts do not map letters back to characters correctly	Try PDF to text; if the result is still garbled, use OCR
The viewer blocks selection	A copy restriction the author set	See the section on restrictions

Why a PDF will not let you copy

Three things stop copying, and each has a different answer:

• The pages are scanned images. The text is not in the file at all, so OCR is the only way to get it back.
• The file needs a password to open. It cannot be read until you unlock it, with the password you have, in the program you normally use for it. The tools here do not try to open locked files.
• The author set a copy restriction. Some PDFs tell viewers not to allow copying or editing. That is a setting the author chose, and fixmint does not get around it. If you have the right to, ask for an unrestricted copy or export one from the program that made the file.

Get the text from a text PDF

1. Drop the PDF onto the box in PDF to text and wait while each page is read.
2. Check the page and word counts that appear.
3. Tick Page markers if you want each page labelled. Leave it off for plain text.
4. Copy the text, or download it as a .txt file.

Columns, tables and headers can come out in a different order, because a PDF stores where each piece of text is drawn, not which paragraph it belongs to. If every line ends with a hard return, read why text copied from a PDF breaks at every line, then paste the text into Fix pasted text, which joins the lines of each paragraph.

Get the text from a scanned PDF

Scanned pages take two steps: draw each page as an image, then read the text in each image. The same reading step works for a screenshot or photo, as explained here.

1. Drop the PDF onto PDF to JPG. Choose PNG for text, because it keeps letter edges sharp, and 300 DPI for small print. Pick only the pages you need, or leave it on all pages.
2. Download the pages. Each is saved as its own image, or all of them come as a ZIP.
3. Open Image to text and drop the first page. The words appear once the page is read. Words the engine is unsure of are marked in red.
4. Repeat for each page in order. Tick Join lines into paragraphs to remove the line breaks, then copy the result.
5. Check every number, name and date against the page before you rely on it.

Image to text reads one image at a time, and only English. The tool loads the OCR engine and its English model from the site when you start, and the image is then read on your device. A large or slow page can take a while, so wait for the result before you copy. Sharp, straight, evenly lit pages read best. Crop out the edges of a photo, and hold the camera square to the page.

Common mistakes

• Saving text pages as JPG. JPG blurs the edges of letters, which costs accuracy. PNG is the better choice for text-only pages.
• Trusting the OCR without a check. Recognition is not perfect. Read the numbers and names against the page.
• Expecting tables to stay tables. OCR reads lines, so a table can come out as a list. Compare the result with the page.
• Using OCR for handwriting. Image to text is built for printed text. Handwriting usually reads poorly, so retype those parts by hand.

Follow-up questions

Does OCR upload my scanned pages?

No. The page is read in your browser, and the image is never sent to a server. The site delivers the OCR engine and its English model to your browser, and the reading happens there.

What if only some pages have no text?

Turn on Page markers in PDF to text. A page that shows nothing under its marker is a scan, while the typed pages show their text. Run only those pages through PDF to JPG and Image to text, then put the results back in order.

Why does the copied text break at the end of every line?

A PDF stores each printed line separately, so copying keeps those breaks. Use Fix pasted text to join the lines of each paragraph. The line-break guide explains why it happens.

Can Image to text read handwriting or other languages?

It is built for printed English in the Latin alphabet, and handwriting usually reads poorly. Other alphabets, such as Arabic, Chinese or Cyrillic, are not supported.