· Guides
Why Urdu Nastaliq Is Hard for OCR (With Real Results)
I ran the same Urdu text in Naskh and Nastaliq through browser OCR. Naskh read perfectly, Nastaliq lost about one letter in five. Results, causes and tips.
Get text out of soft, grainy or low contrast photos. The image is cleaned up before reading, and a confidence score tells you how far to trust the result.
Supports JPG, JPEG, PNG, GIF, BMP, WebP, TIFF, HEIC, HEIF, HEICS
Auto detects the writing system and reads it. For better results, pick the language (up to 3 if the text mixes languages).
Drop a blurry photo, screenshot or scan here
or
Max file size: 10MB
Convert different image formats to text with our specialized OCR tools
Add the image, try the black and white option if the text is faint, and check the confidence score.
Last updated:
1
Drop or choose the blurry photo, screenshot or scan (up to 10 MB). Crop it to the text first if you can.
2
Click Extract Text. The image is converted to grayscale, its contrast is stretched and small text is enlarged. For faint text, tick Black and white and try again.
3
Look at the confidence score, fix any wrong words, then copy the text or download it as a .txt file.
I set a library opening hours notice in Helvetica, then added a soft blur and a small sideways shake, like a photo taken with a moving hand, plus grain and JPEG compression. A person can still read it. Here is what the tool made of it.
Extracted text with English picked, unedited
tk yo. air how Monday to Thursday: 9:00 t0 20:00 Friday: $200 t 17:00 Samsday. 10:00 & 16:00 Sunday. closed Returns can‘ . taf in the box by the side door.
The tool reported: Read as English. Confidence: 64%. OCR language: English, picked by hand.
These steps run before the text is read. None of them can sharpen detail that is not there, but they make the most of what is.
Faded, grey or washed-out text is spread across the full range from black to white, so letters stand out from the background.
Uneven light and shadows across a photographed page are balanced before reading.
When lines of text are small, the image is upscaled before reading, which helps soft, low resolution pictures.
Turns the image into pure black and white. It often rescues faint print, and sometimes hurts photos with detail: try both.
Auto mode also tries your browser’s language, which helps when blur makes the writing system hard to detect.
You see how sure the OCR is, so you know when to trust the text and when to retake the photo.
Indoor pictures of documents taken without enough light.
Photos taken quickly, on the move or with a shaking hand.
Pictures forwarded through messaging apps that have lost quality.
Grainy, low resolution scans and photocopies of photocopies.
Screenshots enlarged until the letters became soft.
Text on signs and boards photographed from far away.
Our image to text converter is designed to streamline the process of extracting text from images. However, we understand you might have a few questions. To make the most of this tool, we've compiled some frequently asked questions along with helpful explanations to guide you.
Open this page, add the image and click Extract Text. If the result is poor, tick Black and white and try again, and crop the image to the text.
No. It improves contrast, lighting and size before reading, but it does not sharpen or reconstruct blurred letters. If you cannot read the text yourself, OCR usually cannot either.
Blurred letters look alike (rn and m, 0 and O, 1 and l), so OCR confuses them. Check the confidence score, correct obvious mistakes, and retake the photo if you can. More on what limits OCR accuracy.
No. It helps most with faint or grey text on plain paper. With photos that have gradients or colored backgrounds it can remove detail, so compare both results.
Small text suffers most from blur. See Small Text OCR for tips on getting tiny print readable.
No. The cleanup and OCR run in your browser. The image is not sent to a server.
Use more light, tap the screen to focus on the text, hold the phone steady for a second, and fill the frame with the page. Our guide to cleaner OCR from photos has more.
Tips, guides, and insights on OCR technology
· Guides
I ran the same Urdu text in Naskh and Nastaliq through browser OCR. Naskh read perfectly, Nastaliq lost about one letter in five. Results, causes and tips.
· Guides
Is photo to text AI? Yes, a small neural network reads each line. How Tesseract's LSTM works, how it differs from AI chatbots, and what that means for you.
· Guides
Image to text with Google Lens or a browser OCR tool? A fair comparison: camera, translation and handwriting vs batch files, downloads and no upload.