To extract text from image files such as photos, screenshots and scanned pages, you need OCR, which reads the letters in the picture and gives them back as text you can copy. Add the image to PDFNova OCR, choose the language, run recognition, then copy the text or download it as a .txt file. It works in the browser on a phone or computer and the picture stays on your device.
This is useful more often than you might expect: a photo of a notice board, a screenshot of a message, a slide from a lecture, a page of a printed form, the label on a product. Typing these out by hand is slow and error-prone. Recognition does most of the work and leaves you to check the result.
Images that work well and images that do not
| Image type | Typical result | Tip |
|---|---|---|
| Screenshot of text on a screen | Very good | Zoom in before taking the screenshot so the letters are larger |
| Clear photo of a printed page | Good | Shoot from directly above in even light |
| Photo of a sign or poster | Fair to good | Stand square to it and crop away the surroundings |
| Photo of a computer or TV screen | Fair | Screen patterns and reflections cause errors, a screenshot is better |
| Text over a busy picture or pattern | Poor | Expect missing words |
| Handwriting | Poor to fair | Only neat, separated letters have a chance |
The pattern is simple. Dark, sharp, straight letters on a plain light background are read well. Anything that makes the letters harder for you to read at a glance makes them harder for software too.
How to extract text from an image with PDFNova
- Open the OCR tool in your browser.
- Add your image. You can also add a scanned PDF in the same way.
- Select the language that appears in the picture. The options are English, Urdu, Arabic, Hindi, French, Spanish, German and Chinese (Simplified).
- Start recognition. The first time a language is used, its data is downloaded once, so that run takes longer than later ones.
- Review the recognised text on screen.
- Choose what to do with it: copy the text, download a .txt file, or download a searchable PDF that keeps the picture with a text layer behind it.
Paste the text into your notes or document and read it once against the image. Pay special attention to numbers, email addresses and names, where a single wrong character matters.
Prepare the picture for better recognition
A minute of preparation often saves ten minutes of correcting.
- Crop to the text. Remove the desk, fingers, browser toolbars and anything else that is not the text you want.
- Straighten it. Rotate the image so lines of text run level. Text on its side is rarely read correctly.
- Keep the original size. Images forwarded through chat apps are often shrunk. Use the original photo from your gallery when you can.
- Retake a blurry photo. No setting will rescue letters that are smeared. Hold still, tap to focus and shoot again.
- Avoid flash on shiny paper. A white glare patch wipes out the words beneath it.
If the text is still on paper in front of you, do not take an ordinary photo at all. Capture it with the Camera Scanner, which detects the page edges, corrects the perspective and offers a black and white filter. That gives OCR a flat, high-contrast page to read. For a full list of improvements, see our tips on how to improve OCR accuracy.
Working with several images at once
When a document is spread over many photos, such as eight pictures of an eight-page letter, handling them one by one gets tedious. Combine them first with Image to PDF, where you can put the pictures in order and choose the page size. Then run OCR on the single PDF and download one text file or one searchable PDF for the whole document.
This also leaves you with a tidy file to archive, instead of loose pictures scattered through your gallery.
Text in Arabic, Urdu, Hindi and other languages
Always match the language setting to the image. Recognition relies on knowing which script and which words to expect, so an Arabic page read with the English setting returns meaningless characters.
Accuracy differs between scripts. Clean printed English is usually the most reliable. Printed Arabic and Hindi work but need closer checking, and there is more detail in our guide to Arabic OCR online. Urdu in the Nastaliq style is harder than printed English because the words slope and stack, so plan to correct more. When an image mixes two languages, run it with the main language and fix the other parts by hand.
What to do with the extracted text
- Paste it into a message or email when you only need to share the words.
- Save the .txt file for notes and study material you want to edit later.
- Keep the searchable PDF for records, so you have the original look plus working search.
- Move it into a word processor when you need a formatted document. Plain text carries no fonts, tables or columns, so you rebuild the layout yourself. The process is explained in how to convert a scanned PDF to Word using OCR.
Troubleshooting image to text problems
- The output is gibberish. Wrong language selected, or the image is rotated.
- Whole lines are missing. The text is too small, too faint or sits on a colored or patterned background. Crop closer and retake at higher quality.
- Letters are confused. Low resolution makes similar shapes hard to tell apart. Use the original image, not a compressed copy.
- Columns are mixed together. Crop the image into one column at a time and process each part.
- Recognition is slow. The work runs on your device, and big photos take longer on older phones. Cropping to the text area helps.
Privacy when converting pictures to text
Photos of documents often contain personal details: addresses, account numbers, ID numbers, private messages. PDFNova processes files in your browser, so your documents are not uploaded to a server. PDFNova is free, needs no sign-up and adds no watermark.
Frequently asked questions
How can I extract text from an image for free?
Use a browser OCR tool. Add the image, pick the language, run recognition and copy the text or download it as a .txt file.
Can I extract text from a screenshot?
Yes, and screenshots usually give the best results because the letters are sharp and the background is clean.
Can I copy text from a photo on my phone?
Yes. Open the OCR tool in your phone browser, choose the photo from your gallery and copy the recognised text into any app.
Does it work with handwritten text in a picture?
Only partly. OCR is designed mainly for print, so handwriting, especially joined writing, produces many errors.
Pick a clear picture with printed text, open PDFNova OCR and try it. In most cases you will have usable text faster than you could type the first paragraph.