Quick answer
To convert an image to text, drop a JPG, PNG or screenshot into the tool (or paste it with Ctrl+V), choose the language of the text and wait a few seconds. The recognised text appears in an editable box — copy it or download it as a .txt file. The image is read by your own browser, so it is never uploaded.
- Input:
- JPG, PNG, WebP, BMP, screenshots
- Languages:
- English, Bengali, Hindi, Arabic, Urdu and more
- Paste support:
- Ctrl+V a screenshot directly
- Output:
- Editable text, copy or .txt download
- Accuracy:
- About 95–99% on clear printed text
- Privacy:
- Image processed on your device
- Handwriting:
- Neat print only
- Sign-up:
- Not needed
Text trapped inside an image is useless for searching, editing or translating. Maybe it is a screenshot of an error message you want to google, a photo of a printed notice, a page from a textbook you need to quote, a business card, a recipe, or an old letter. Retyping it is slow and introduces mistakes. OCR — optical character recognition — reads the letters in the picture and turns them into real text you can copy.
Our image to text converter uses Tesseract, the open-source OCR engine originally developed at HP and later improved by Google, compiled to run inside your web browser. It recognises printed text in more than 100 languages, including complex scripts like Bengali (বাংলা), Hindi, Arabic and Urdu, and mixed English–Bengali documents. Each language model downloads once and is cached, so later conversions start quickly.
Because the recognition happens on your device, the image is never sent to a server. That is important when you extract text from ID cards, bank documents, private chats or medical letters. Results appear in an editable box, so you can fix the occasional misread character before copying the text into Word, Google Docs, an email or a translation tool.
How to extract text from an image
- 1 Pick the languageChoose the language of the text in the image, such as English, Bengali or "English + Bengali" for mixed documents. The right language improves accuracy dramatically.
- 2 Add the imageDrop a photo or screenshot onto the tool, click to browse, or simply press Ctrl+V (Cmd+V on Mac) to paste a screenshot you just took.
- 3 Wait for recognitionThe OCR engine reads the image — usually 2–10 seconds. A progress bar shows how far it has got.
- 4 Check and correctRead through the text in the box. The confidence score tells you how sure the engine is; fix any wrong characters directly in the box.
- 5 Copy or downloadClick Copy to paste the text anywhere, or Download .txt to save it as a file.
Why use this OCR tool?
Many languages
English, Bengali, Hindi, Arabic, Urdu, Spanish, French, German, Chinese, Japanese and more.
Paste screenshots
No need to save a file first — paste a screenshot straight from the clipboard.
Editable results
Correct mistakes before you copy, and download the text as a .txt file.
Confidence score
See how certain the engine is, so you know when to double-check.
Private
The image is read on your device and never uploaded.
Free
Use it as often as you like, without an account.
What kinds of images work best?
| Image | Expected accuracy | Tip |
|---|---|---|
| Screenshots of websites, apps, PDFs | Excellent (98–99%) | Capture at 100% zoom or larger |
| Scanned printed pages | Excellent | Scan at 300 DPI, black and white |
| Phone photos of documents | Very good | Shoot straight from above in good light |
| Signs, posters, packaging | Good | Crop to the text area first |
| Neat handwriting in print letters | Fair | Dark pen on white paper |
| Cursive handwriting | Poor | Not reliably supported |
How to get better OCR results
- Choose the correct language. Using the English model on Bengali text produces gibberish.
- Crop to the text. Remove photos, logos and decorations around the text with our Image Cropper.
- Straighten the photo. Tilted lines confuse OCR. Take the photo square-on, or rotate it before converting.
- Use enough resolution. Letters should be at least 20 pixels tall. Zoom in before taking a screenshot of small text, or enlarge small images with our Image Upscaler.
- Good contrast. Dark text on a light background works best. Avoid shadows and glare on glossy paper.
OCR for Bengali and other Indian languages
Bengali is one of the hardest scripts for OCR because of conjunct letters (যুক্তবর্ণ), vowel signs placed around consonants and the connecting headline (মাত্রা). Tesseract's Bengali model handles printed Bengali well, especially newspapers, books and government notices in common fonts such as SolaimanLipi, Kalpurush and Nikosh. For documents mixing English and Bengali — for example forms and bills — choose "English + Bengali" so both scripts are recognised in one pass. Hindi (Devanagari), Urdu and Arabic are also supported.
Common uses for image-to-text
- Copying text from screenshots of chats, error messages or social media posts.
- Digitising printed notes, handouts and book pages for study.
- Extracting details from business cards, receipts and invoices.
- Translating signs and menus when travelling — paste the text into a translator.
- Making accessible text versions of images for visually impaired readers.
- Searching old photographed documents by turning them into text.
Images vs. scanned PDFs
This tool reads one image at a time. If you have a multi-page scanned PDF, use our PDF OCR tool instead: it reads every page and can create a searchable PDF where the original scan stays visible but the text can be selected, copied and searched. To pull pages out of a PDF as images first, use PDF to PNG.
How OCR works
OCR first cleans the image — converting it to black and white and removing noise. It then finds blocks of text, splits them into lines, words and individual characters, and compares their shapes with patterns learned from millions of examples. Modern engines like Tesseract 5 use an LSTM neural network that reads whole lines at once, using context to decide between similar characters such as "0" and "O" or "l" and "1". The result is plain text plus a confidence value for each word, which is where the confidence score shown under the box comes from.
Built-in alternatives on phones and computers
Modern devices have some OCR built in, and it is worth knowing when each option fits:
- iPhone (Live Text): press and hold on text in a photo to select it. Great for quick English copies; limited language support.
- Android (Google Lens): point the camera at text and tap "Text". Works well but sends images to Google.
- Windows (Snipping Tool, PowerToys Text Extractor): handy for screenshots in common languages.
- Mac (Live Text in Preview): select text in images directly.
Is online OCR private?
Many OCR websites upload your picture to their servers, and some keep it to train their models. Here, the OCR engine itself is downloaded into your browser, and recognition runs locally on your processor. Your image and the extracted text stay on your device — we only count that the tool was used, to apply daily plan limits. You can even disconnect from the internet after the language data has loaded and the tool keeps working.
Troubleshooting
- The result is gibberish. Check the language setting and make sure the image is not upside down.
- Some letters are wrong. Low resolution or blur is the usual cause. Retake the photo closer and in better light.
- Language data loads slowly. Each language downloads once (a few MB) and is then cached in your browser.
- Columns are mixed up. Crop each column separately for magazine or newspaper layouts.
Frequently asked questions
How do I copy text from an image for free?
Drop the image into the tool, choose the language and wait a few seconds. The text appears in a box where you can copy it or download it as a .txt file.
Can it read Bengali text?
Yes. Choose Bengali, or English + Bengali for mixed documents. Printed Bengali in common fonts is recognised well.
Can it read handwriting?
Neat handwriting in separate print letters can work, but cursive handwriting is not reliably recognised.
Is my image uploaded?
No. The OCR engine runs in your browser and the image stays on your device.
How accurate is the OCR?
Clear screenshots and scans of printed text usually reach 95–99% accuracy. Blurry or tilted photos produce more errors.
Can I paste a screenshot directly?
Yes. Take a screenshot, open this page and press Ctrl+V (or Cmd+V on Mac).
Can I convert a scanned PDF to text?
Use our PDF OCR tool, which reads every page of a PDF and can create a searchable PDF or a text file.
Which image formats are supported?
JPG, PNG, WebP, BMP and GIF images, plus screenshots from the clipboard.
Does it work on mobile phones?
Yes. You can take a photo with your phone's camera from the upload button and extract the text right away.
Can I extract text from several images at once?
Process them one after another — each image takes only a few seconds, with no daily limit.
Last updated November 22, 2026 · Written and tested by the SmallEasyTools team.