Can't copy text from an image? Here's how to extract it
Quick answer
Upload the image to the OCR tool, and the text is read and returned as real, selectable text you can copy or download - typically within a few seconds.
- 1Upload your image.
- 2Choose Image to Text.
- 3Wait for recognition to finish.
- 4Copy the text or download it as a file.
Free, no sign-up, and your file is read on your own device rather than uploaded.
An image is just a picture, even when it is a picture of text - which is why trying to select or copy text out of a JPG or PNG does nothing. There is no text there for your browser to select, only pixels that happen to look like letters.
OCR (optical character recognition) reads those pixels and works out what letters and words they actually form, turning the picture back into real, selectable, copyable text. This runs entirely in your browser, so the image never leaves your device.
Step by step
Wait for recognition
A progress indicator shows what is happening, from initializing to recognizing the text itself.
OCR Text ExtractorCopy or download the result
Copy the text directly to your clipboard, or download it as a .txt file.
OCR Text Extractor
Tips
- A clear, well-lit, straight-on photo of text gives the most accurate result - heavy glare, extreme angles or very low resolution all make recognition harder.
- If the image is rotated or upside down, the result will be inaccurate - rotate the image the right way up before uploading it.
- This is built for printed or typed text, not handwriting - handwritten notes will not be recognised reliably.
Common problems
The extracted text has a few wrong characters here and there.
This usually comes from image quality - blur, low resolution or heavy compression. A sharper or higher-resolution version of the same image generally fixes it.
The result is nonsense or unrelated to the image.
Check whether the image is rotated or sideways - that alone can produce this. Straighten the image first and try again.
It couldn't read handwritten notes.
This OCR engine is built for printed and typed text, not handwriting, which needs a different kind of model entirely.
Frequently asked questions
- Why can I not just select the text in my image?
- An image has no real text in it, only pixels - even when those pixels are shaped like letters. Selecting text only works on genuine text characters, which is exactly what OCR creates from the picture.
- Is my image uploaded to a server?
- No. Recognition happens entirely in your browser, on your own device.
- What image formats are supported?
- JPG, PNG and WebP.
- Can it read handwriting?
- Not reliably - this engine is trained on printed and typed text.
- How accurate is the extracted text?
- Very accurate for a clear, well-lit photo or screenshot of printed text. Accuracy drops with blur, glare, very low resolution or an unusual font.
Related OCR tools
Related articles
- OCR
Need text from a screenshot? Extract it in your browser
Upload the screenshot to the OCR tool. Screenshots usually convert very accurately, since the text in them is typically crisp and high-contrast rather than photographed.
- OCR
How to convert an image to text
Upload the JPG, PNG or WebP file to the OCR tool and it reads the text in the image, returning it as real text you can copy or download as a .txt file.
- OCR
How to extract text from a scanned PDF
Upload the scanned PDF to the OCR tool and choose PDF to Text. Each page is read in turn, and any page that already has real text is used directly rather than re-read - only pages that are genuinely scans get OCR.
Ready to do it?
Free, no sign-up, and nothing is uploaded to a server.