How to Extract Text from a Scanned PDF

To extract text from a scanned PDF, open the Image to Text tool, choose the PDF, pick the language and tap Read the text. The tool reads each page as a picture and gives you text you can copy or download as a .txt file.

A scanned PDF holds pictures of pages, not real text, which is why you cannot select the words. Reading pictures into text is called OCR. It works well on clean scans, and less well on blurry or handwritten ones.

What you need

• The scanned PDF
• The language of the text. English, Hindi, Malayalam and Arabic are supported.
• A little patience. Each page is read one after another, and a long PDF takes a while on a phone.

Steps to get text from a scanned PDF

1. Open the Image to Text tool, tap the box and choose your PDF. The file stays on your device. It is not uploaded.
2. Choose the language of the text under Language of the text. The size of the language data is shown before anything is downloaded. English is already chosen.
3. Tap Read the text. The first time, the reading engine and the language data download once. After that they are saved on your device.
4. Wait while the tool reads each page. A message shows which page it is on.
5. Read the result in the text box. You can correct mistakes right there.
6. Tap Copy text to copy it, or download it as a .txt file.
7. Check numbers, names and dates against the original. This is the part that most often has mistakes.

How well does it work?

The result depends on the scan. These things give good results:

• Printed text, not handwriting
• A sharp, straight scan, with the page flat
• Dark text on a plain light background
• A common font in a normal size

These things cause mistakes:

• Blurry or low-quality scans
• Shadows, glare or creases across the text
• Handwriting, stamps or decorative fonts
• Text over a coloured or patterned background
• A page that is sideways or tilted

Always read the result before you rely on it. Treat it as a fast first draft, not a perfect copy.

Get a better result

If the first try is poor, change one thing at a time:

1. Fix the direction. If pages are sideways, rotate them first with Rotate PDF. Text read from a sideways page is mostly nonsense.
2. Trim the edges. Margins with dark scanner shadows can confuse the reading. Trim them with Crop PDF.
3. Choose the right language. The wrong language gives strange letters. Hindi, Malayalam and Arabic options also read English words and numbers on the same page.
4. Use a better scan. If you can scan again, use even light, hold the phone steady and fill the frame. Our guide to scanning documents with your phone explains how.
5. Work page by page. Use PDF to JPG to turn the pages into pictures, then read the best ones.

What you get, and what you do not

You get plain text. The tool does not rebuild the layout, tables or columns of the page, and it does not make a searchable PDF. If you need text from a table, check the arrangement by eye after copying it.

For a single photo or a screenshot rather than a PDF, see our guide on copying text from a photo.

Common problems

The result is full of strange symbols. The scan may be poor, the page may be sideways, or the language may be wrong. Fix the direction and the language and try again.

Some words are missing. Text over dark or patterned areas is often skipped. Try a cleaner scan.

It is slow. Each page takes a few seconds or more on a phone. Keep the tab open until it finishes.

The first read took longer than the next. The engine and language data download once the first time. After that they are saved on your device, and later readings start straight away.

Frequently asked questions

Why can I not select text in my PDF?

Because the PDF holds pictures of pages, for example from a scanner. There is no real text in it. OCR reads the pictures and turns the words into text.

Can it read Hindi or Malayalam?

Yes. English, Hindi, Malayalam and Arabic are supported. Each language’s data is downloaded only when you choose it, and the size is shown first.

Is my PDF uploaded?

No. The reading happens in your browser on your own device. Your PDF and the text never go to our servers or anyone else.

Will the layout stay the same?

No. You get plain text. Columns, tables and pictures are not rebuilt, so check the result against the original.

How accurate is it?

It depends on the scan. A clean printed page is usually read well. Blur, handwriting and glare cause mistakes, so always read the result before you use it.