Skip to content
EN
All free toolsMenu

Type a task or a name, for example “remove background”, “PDF” or “P0420”.

Documents·7 min read

Scanned image to text: get editable words from paper and PDFs

A letter from the bank, an old recipe, a contract that came back as a scan. You want to quote it, search it or fix a typo, but the page is only a picture. Here is how to turn it into text, including when the scan is a PDF.

By LoschaZ·Updated

The short answer

First check whether the PDF already has text: try to select a single word. If that works, just copy it. If not, save the page as a picture (a screenshot is fine), open it in the AI Text Scanner, pick the language and copy the result. A straight, sharp scan with dark text on a light page reads best.

Open the free tool

Does your scanned PDF already contain text?

It might. Some scan apps and PDF programs run text recognition themselves and hide the text behind the picture of the page. Check this first, because then there is nothing to convert.

  1. Try to select one word

    Open the PDF in the reader you normally use. On a computer, drag across a single word. On a phone, touch and hold a word.

  2. Look at what gets selected

    If just that word is highlighted, the PDF has text: copy it and you are done. If the whole page is selected like a photo, or nothing happens, the page is only a picture.

  3. Search as a second check

    Press Ctrl+F (Cmd+F on a Mac) and search for a word you can see on the page. If the reader finds it, there is text behind the picture.

How to change a PDF image to text: save the page as a picture first

Save the page as a picture, because the AI Text Scanner reads images, not PDF files. You can do it with the reader and the computer or phone you already have.

The easiest way is a screenshot. Zoom in until the page fills the screen and the letters are big, then take it. Big letters on screen give a better result: Google’s help for its own text recognition asks for text at least 10 pixels high.

  • Windows: press the Windows logo key + Shift + S and drag over the page.
  • Mac: press Shift-Command-4 and drag over the page. The screenshot lands on your desktop. Add Control (Control-Shift-Command-4) to copy it to the clipboard instead.
  • iPhone: press the side button and the volume up button at the same time. On an iPhone with a Home button, press the side button and the Home button.
  • Android: press the Power and Volume down buttons at the same time. Some phones use other buttons; the phone maker’s help explains them.
  • Mac Preview can also save a page as a picture: open the PDF, choose “File” > “Export”, and pick PNG or JPEG in the “Format” menu.

How to scan a picture to text with a clear result

Scan the page straight, sharp and evenly lit. That decides more about the result than any setting. A flatbed scanner makes it easy, but a phone works too.

A flatbed scanner with its lid open. A printed page lies on the glass, lined up with the edge, and a light bar glows along one side.
A flatbed scanner keeps the page flat and evenly lit. Line the page up with the edge of the glass, and the lines of text come out straight.

  1. Lay the page flat

    Put the page face down on the glass, lined up with the edge, and close the lid. Smooth out folds. With a book, press the page flat near the spine.

  2. Pick the resolution

    Choose 300 dpi if the scanner asks. The Tesseract documentation says its engine, which our scanner uses, works best on images of at least 300 dpi.

  3. Save it as a picture

    Save the scan as PNG or JPG. If the scanner software only offers PDF, save the page as a picture as shown above.

How to convert a scanned image to text, step by step

Open the picture page of the AI Text Scanner and add your scans. There, “Boost contrast” and “Enlarge small images” under “Prepare the image” are already switched on, which helps with scans and phone photos.

  1. Add the pages

    Drag the scans onto the page or press “Choose images”. On a phone, “Take a photo” opens the camera, and “Choose images” opens your gallery.

  2. Set the language

    Under “Text language”, pick the language of the document: English, German, French or Spanish.

  3. Turn sideways pages

    If a page lies on its side, use “Rotate left” or “Rotate right”. To read only one part, such as a single column, drag across it; on a touch screen, tap “Select part” first.

  4. Read, check, then copy or save

    Press “Read text”. Compare the result with the page and correct it in the text box. Turn on “Join broken lines” to put sentences that wrap across lines back together. Then press “Copy text” or “Download .txt”.

Why does a scanned image give the wrong text?

Usually because of the scan: a tilted page, a shadow, blur, small print or a colored background. Each one has a fix.

Two pages side by side. On the left, a straight, sharp page with even lines of text and a green check mark. On the right, the same page tilted, blurry and shadowed, with a red cross.
The same page, scanned twice. The straight, sharp scan on the left reads well. The tilted, shadowed, blurry scan on the right gives broken words.
  • Tilted lines: scan again with the page straight. The Tesseract documentation says the quality drops a lot when a page is too skewed. “Rotate left” and “Rotate right” turn a page by 90°, so they fix sideways pages, not slight tilts.
  • Shadows and gray patches: scan again in even light, or turn on “Boost contrast”. Tesseract names an unevenly dark background as a cause of poor results.
  • Small print: scan at a higher resolution, or turn on “Enlarge small images”.
  • Two columns: the lines of both columns can get mixed up. Select one column at a time.
  • Tables: they come out as plain lines of text. Rebuild the table by hand if you need it.
  • Stamps, logos and colored paper: crop them out, or select only the text.

Handwriting picture to text: what to expect

Neat handwriting in separate letters sometimes works. Joined-up writing, quick notes and signatures usually do not.

The Tesseract project, whose engine our scanner uses, says plainly that it is designed for printed text and will not work very well on handwriting. So treat any result as a rough draft.

  • Choose pages with clear block letters in dark ink.
  • Photograph them in good light, straight from above.
  • Check every word against the paper. Names and numbers need the most care.
  • For a long handwritten letter, typing it yourself may be faster than correcting the result.

Questions you might still have

Can I turn a scanned PDF into a Word document?

You can get the words into Word, but not the layout. Copy the text from the scanner, or download it as a .txt file, and paste it into Word or Google Docs. Bold text, columns and tables have to be redone by hand.

Can I convert a whole multi-page PDF at once?

Not with the AI Text Scanner, because it does not open PDFs: save each page as a picture and add them all at once. Google Drive can open a PDF with Google Docs and convert the text, but the file is uploaded to your Google account, Google asks for files of 2 MB or less, and lists, tables and columns are not likely to be detected.

What resolution should I scan at?

300 dpi is a good choice: the Tesseract documentation says its engine works best on images of at least 300 dpi. For a phone photo, fill the screen with the page and check that the letters stay sharp when you zoom in.

Which languages can the scanner read?

English, German, French and Spanish, one at a time. For a page in two languages, read it twice, once with each language, or select each part separately.

Is it safe to scan private documents?

With the AI Text Scanner, the picture stays on your device. It reads in your browser and does not store or send the image or the text. Your browser only keeps the language file, so it does not download again.

Why are some words split with a hyphen?

The page split them at the end of a line, and the scanner copies the line breaks. “Join broken lines” joins the lines inside a paragraph and repairs words split by a hyphen at a line end, but keeps the paragraph breaks.

Sources & useful details

These guides are written from the current LoschaZ tools and the sources listed above. Examples are illustrations, not tests on every device or promises of an identical result.

What to try next

All tips & guides