How to Extract Text From an Image or Scan (Free OCR for Bangla and English)
Copy text out of a photo, screenshot, scan or certificate without retyping. Learn how OCR works and how to get the cleanest result.
You have a photo of a notice, a scanned page, a screenshot of a message or a picture of a certificate. You need the text inside it, to edit, to search or to paste into another document. Retyping is slow and full of mistakes, especially in Bangla. OCR solves this by reading the letters in the picture for you.
The free Image to Text tool reads Bangla, English, Hindi and Arabic text and gives you a plain text file you can copy.
Quick answer
Open Image to Text, choose your picture, select the language of the text, press the button and download or copy the text. Clear, straight, well-lit pictures give the best result.
What is OCR?
OCR stands for optical character recognition. The software looks at the shapes in a picture and decides which letters they are, turning an image of text into real text that you can select, copy and edit. It works well on printed text and less well on handwriting.
How to extract text, step by step
- Open the Image to Text tool.
- Add your picture or scan.
- Choose the language that matches the text, for example Bangla or English.
- Press the extract button and wait for the result.
- Copy the text or download it as a
.txtfile, then paste it into Word or a message.
How to get a more accurate result
The quality of the picture decides the quality of the text. Follow these rules:
- Keep the paper straight. Photograph from directly above, not at an angle. If a phone photo is skewed, clean it first with the Document Scanner.
- Use even light. Avoid shadows and glare across the page.
- Use a large, sharp picture. Blurry or tiny text is read badly.
- Choose the right language. A Bangla page read as English gives nonsense.
- Crop to the text if the picture has a busy background.
What works well and what does not
| Works well | Needs checking |
|---|---|
| Printed books, letters and notices | Handwriting |
| Typed certificates and forms | Very small or faint print |
| Screenshots with clear text | Photos with heavy shadows |
| Clean scans | Decorative or stylised fonts |
Always read the result once, particularly numbers, names and dates, before you rely on it.
Scanned PDFs and Word documents
If your source is a scanned PDF rather than a picture, you can turn it into an editable document with the Image / Scan to Word (OCR) tool, which produces a Word file. For PDFs made from real text, use PDF to Word instead, because there is no need for OCR.
Typical uses
- Pulling an ID number or address from a card for a form
- Digitising a book page or class notes
- Copying text from a screenshot of a chat or a poster
- Making old printed documents searchable
- Translating text by pasting it into a translation tool
Privacy
Your picture is used only to read the text, and then it is deleted. No account is needed and nothing is stored in a database. Be careful with highly sensitive documents on any online service, and read the privacy policy.
Real situations where OCR helps
Digitising old documents. Typing a ten-page letter takes an hour; OCR does it in seconds and you only proofread.
Forms that require typed information. Copy a name or number from a photographed ID or certificate rather than retyping and risking a mistake.
Studying from photographed notes. Turn a photo of a handout into searchable text.
Collecting information from screenshots, such as a chat, a poster or a bank message.
How to prepare the picture for the best accuracy
Accuracy depends on the picture more than on the software:
- Scan or photograph at a good resolution, at least about 300 DPI equivalent for small print.
- Straighten the page. Skewed lines confuse recognition.
- Use even lighting without shadows or reflections.
- Increase contrast if the paper is grey or the print faint.
- Crop away borders and unrelated pictures.
- Select the right language. For mixed Bangla and English, choose the combined option if available.
Proofreading the output
OCR is accurate on clean printing but never perfect. Check these frequent problems:
- Numbers and dates, where 0 and O, 1 and l are mixed up.
- Names and addresses, which a dictionary cannot correct.
- Bangla conjuncts, which may be split or merged.
- Line breaks, which may split a sentence in the middle.
- Tables, which usually lose their columns in plain text.
Read the result side by side with the original, and especially verify anything that will be used in an official form.
What to do after extracting
- Paste into Word and format it, then export with Word to PDF.
- Count words with the Word Counter.
- Convert old Bijoy text with Bijoy to Unicode when the source used that encoding.
- Make a scanned PDF searchable with OCR PDF.
Quick checklist for better OCR
- Use a sharp, well-lit and straight picture.
- Crop to the text.
- Choose the right language.
- Proofread numbers and names.
- Save the text in a document with the source picture for reference.
Frequently asked questions
Which languages are supported? Bangla, English, Hindi and Arabic.
Does it read handwriting? Printed text reads well. Handwriting is not always correct, so check the result.
What file do I get? A plain text (.txt) file that you can copy into Word or a message.
How can I improve a poor result? Keep the picture straight and evenly lit, use a large sharp picture and pick the right language.
Is it free? Yes, with no sign-up.
Is OCR accurate for handwriting? Not always. Printed text works well; handwriting should be checked carefully.
Can I extract text from a PDF? For digital PDFs use PDF to Text. For scanned PDFs use OCR.
Which languages are supported? Bangla, English, Hindi and Arabic.
Why are Bangla letters sometimes split? Complex conjuncts are hard to recognise when the picture is small or blurred. Use a larger, sharper image.
Can OCR read text in a photo of a sign? Often yes, if the text is large and clear. Perspective and glare reduce accuracy.
Next step
Open the Image to Text tool below and stop retyping.
