To extract text from an image, open the free Image to Text (OCR) tool, choose the text language, keep “Enhance image before reading” on, and drop your image or screenshot. The text appears in a box you can copy or save as a .txt file. It runs in your browser, so nothing is uploaded.
What is OCR?
OCR stands for optical character recognition. It looks at the shapes in an image and works out which letters and words they are. The result is real text that you can copy, edit, search and translate.
The Img Stop OCR tool uses Tesseract, a well-known open-source OCR engine, running inside your browser. The first time you use a language, the engine and that language’s data are downloaded once. Your images are never sent to a server.
Which languages are supported?
The tool supports 19 options:
- South Asian and Middle Eastern: Urdu, English + Urdu, Arabic, Hindi, Bengali, Persian, Punjabi (Gurmukhi), Turkish.
- European: English, Spanish, French, German, Portuguese, Italian, Russian.
- Asian: Indonesian, Chinese (Simplified), Japanese, Korean.
Choosing the correct language is the single most important setting. OCR set to English cannot read Arabic letters, and the reverse is also true.
How do I extract text from an image step by step?
- Open the Image to Text tool.
- In Text language, choose the language of the text in your image.
- Keep Enhance image before reading (recommended) ticked. It converts the image to grayscale, increases contrast and enlarges small images before reading.
- Drop one or more images or screenshots. You can also paste a screenshot with Ctrl+V.
- Wait while the tool reads the text. A progress percentage is shown.
- Check the result. The card shows a confidence score and word count. Click Copy or download the .txt file.
For Urdu, Arabic and Persian, the result box is set to right-to-left, so the text displays in the correct direction.
How do I get better OCR accuracy?
OCR quality depends mostly on the image you give it. These tips help in every language.
- Crop to the text. Remove borders, photos, logos and other text blocks with the Crop Image tool. Less clutter means fewer mistakes.
- Keep the page straight. Photograph documents from directly above. Tilted or curved lines are harder to read.
- Use good, even light. Shadows across the page and glare on glossy paper hide letters.
- Make the text big enough. Tiny text in a large photo reads poorly. Get closer or crop tighter.
- Increase contrast if needed. Faded print or pale ink can be improved with the Photo Filters & Adjust tool by raising contrast or using grayscale.
- Use the original image. A screenshot of a screenshot, or a heavily compressed image from a chat, loses detail.
Tips for Urdu OCR
- Pick Urdu, or English + Urdu for mixed text. Many forms, bills and notices mix both scripts. The combined option reads both in one pass.
- Printed Naskh-style text is easier than Nastaliq. Nastaliq, the flowing style used in most Urdu newspapers and books, has stacked and overlapping letters. OCR finds it harder, so expect to proofread more carefully.
- Use larger, clear text. Dots (nuqta) decide many Urdu letters. If the image is small or blurry, dots get lost and letters change.
- Avoid decorative fonts and handwriting. OCR works best on clean printed text.
Tips for Arabic OCR
- Choose Arabic, not Urdu or Persian. The scripts share letters but differ in some characters and spelling, and the right language data gives better word results.
- Diacritics (harakat) are hard. Text with full vowel marks, such as some religious or children’s texts, may come out with missing or misplaced marks. Plain text without marks usually reads better.
- Crop one column at a time. Right-to-left pages with several columns can mix lines. Cropping each column keeps the order correct.
Tips for Hindi OCR
- Choose Hindi for Devanagari text. For Bengali or Punjabi (Gurmukhi), choose those languages instead.
- Keep the top line (shirorekha) sharp. Devanagari letters hang from a line. Blur or a low angle can join or split letters by mistake.
- Watch conjunct letters and matras. Combined letters and vowel signs are the most common errors, so check them when proofreading.
How do I extract text from a PDF?
If a PDF is a scan, its pages are just pictures. First turn each page into an image with the PDF to Image tool, then drop those images into the OCR tool.
What results can I expect?
| Source | Typical result |
|---|---|
| Clear screenshot of printed text | Usually very good |
| Well-lit photo of a printed page | Good, needs light proofreading |
| Nastaliq Urdu or text with many diacritics | Fair, needs careful proofreading |
| Handwriting | Often poor; OCR engines like Tesseract are designed for printed text |
The confidence score helps. A low score is a sign to improve the image and try again.
What can I use OCR for?
- Copying a phone number, address or code from a screenshot.
- Turning a photo of a printed notice or letter into text you can edit or translate.
- Saving quotes from book pages without typing them.
- Digitising old printed documents, receipts and bills.
Key takeaways
- Choose the correct language before reading; use English + Urdu for mixed text.
- Keep “Enhance image before reading” on for most images.
- Crop to the text, shoot straight and use good light for the best accuracy.
- Nastaliq Urdu, Arabic diacritics and Hindi conjuncts need careful proofreading.
- OCR runs in your browser, so private documents stay on your device.
Frequently asked questions
Can OCR read Urdu text from an image?
Yes. Choose Urdu, or English + Urdu for mixed text, in the Text language list. Clear printed text gives the best results. Nastaliq-style Urdu is harder for OCR than simpler printed styles, so always proofread the result before you use it.
Can I extract text from a screenshot on my phone?
Yes. Open the Image to Text tool in your phone browser, choose the language and select the screenshot from your gallery. The text appears in a box below, where you can copy it or download it as a .txt file.
Why is the OCR result full of mistakes?
Common causes are the wrong language setting, small or blurry text, shadows, tilted pages and busy backgrounds. Choose the correct language, crop to the text, keep the enhance option on and use a sharper, well-lit image. Then try again.
Can OCR read handwriting?
OCR engines like Tesseract are designed for printed text. Very neat handwriting may partly work, but results are usually poor. For handwritten notes, expect to correct a lot of words, or type the text yourself if accuracy matters.
Is my document uploaded when I use OCR?
No. On Img Stop, the OCR engine runs inside your browser. The engine and language data download once, and then your images are read on your own device. This makes it suitable for private documents such as bills, letters and forms.