Image to Text (OCR)
Recognize Korean and English text in a picture and copy it as editable text. The image is processed in your browser, not uploaded.
Loading the tool…
User reviews
No reviews yet. If you’ve used this tool, be the first to leave one.
Image to Text (OCR): complete guide
Walk through the steps and examples, or jump straight to the topic you need.
At a glance
- Supported files
- One PNG, JPG or WebP image, up to 20 MB.
- Language
- Choose from Korean, English and Korean + English.
- Processing size
- The long side is reduced to 2,000 pixels for recognition.
- Editing the result
- You can edit the recognized text in place and revert it.
- Confidence
- When recognition finishes, the confidence is shown as a percentage.
- Cancel
- If you cancel while it runs, it stops immediately and you can try again.
- Data source
- Tesseract and its language data are loaded directly from this site. Both are Apache-2.0.
- Where it runs
- The image is not sent to a server and is processed only inside the browser.
How to use Image to Text (OCR)
Add an image with clear text
Images with sharp text, like screenshots, work best. If you photograph paper, shoot it large and straight on, without shadows. Tilted or blurry photos lose a lot of accuracy.
Choose the language to recognize
If you are sure there is only one language, choosing just that language is more accurate. If Korean and English are mixed, use Korean + English. Changing the language clears the previous result.
Watch the progress and wait
The preparation and recognition stages are shown as percentages. The first time takes longer because the recognition engine and language data are downloaded; after that the browser keeps them and it is faster. If it takes too long, click Cancel to stop it immediately.
Fix the result and take it with you
Don't leave the result as is; check the digits and symbols by eye. You can edit right in the result box, and there is also a Revert to recognized text option. Copy it as edited or save it as TXT.
What it’s good for
Copy text out of screenshots and photos of documents
If you captured a screen where copying is blocked, or photographed a paper document, you don't need to retype it. Add the image, choose the language, and the tool shows the text it found, organized line by line. You can edit the result right there, so you only fix a misread digit or symbol and then copy it.
Screens that mix Korean and English
Choosing Korean + English looks for both languages together. If you are sure there is only one language, choosing just that language is more accurate. When recognition finishes, the average confidence is shown too, so you can judge how carefully to check the result.
Worked examples
An order history that mixes languages
900 × 420px image
Korean + English
Receipt total amount
Order number 20260909
Total Amount 12,500 KRW
Thank you for your order
Delivery takes two days5 lines · 92% confidence
Receipt total amount
Order number 20260909
Total Amount 12,500 KRW
Thank you for your order
Delivery takes two daysEditing the result and saving
Change the recognized 12,500 to 12,900The heading shows "Edited by hand"
The saved TXT also contains 12,900How to read confidence
Swipe the table sideways to see the rest.
| Range | Meaning | What to do |
|---|---|---|
| 85% or higher | Mostly read clearly. | Just skim the digits and symbols. |
| 60–84% | Some characters are blurry. | Read through all of it once and fix it. |
| Below 60% | The image is hard to read. | Try again with a larger, clearer photo. |
When the result isn’t what you expected
It says it found no text
This is not a failure; it means no readable text was found. It happens when the text is too small or has weak contrast with the background. Crop out just the text part, enlarge it and add it again.
Hangul comes out as strange characters
If the language is set to English, Hangul is read as similar-looking Latin letters. Switch to Korean or Korean + English and try again.
It is stuck in the preparation stage
The recognition engine or language data could not be downloaded. After a while it reports a failure, and you can reload the page and try again. You can also click Cancel to stop it right away.
Handwriting is not recognized
This feature is built for printed text, so handwriting is outside what it supports. Tables, stamps and vertical writing are not accurate either.
Key terms
- OCR
- Technology that finds text in an image and turns it into text.
- Confidence
- A value showing how sure the tool is about the recognized text. It does not guarantee accuracy.
- Language data
- A file holding the letter shapes of a language. It is downloaded once the first time.
Image to Text (OCR) FAQ
Does it recognize handwriting?
It is built for printed text. Handwriting is mostly not recognized or is badly wrong. Accuracy also drops a lot for blurry or tilted photos and very small text.
Is my photo uploaded to a server?
No. Recognition happens inside this browser. Only the recognition engine and language data are downloaded from this site, and it takes time only the first time.
What if the result looks wrong?
You can edit it right in the result box, and there is also a Revert to recognized text option. Trying again with a larger, clearer photo improves accuracy.
Can it keep the layout of a table?
It carries over only the text, line by line. The cell structure and formatting of a table are not kept. Please tidy it up yourself afterward.
Why is it slow only the first time?
Because the recognition engine and language data are downloaded once the first time. After that the browser keeps them and it gets faster.
Can I add a PDF?
It accepts only one image. For a PDF, extract an image first with the PDF to JPG converter and then add it.