Does this OCR upload my image?
No. The selected image and extracted text are processed in this browser. The page downloads the public OCR engine and selected language model when you start, and the browser may cache those public files.
Tools
Image to Text (OCR)
Extract editable English, Traditional Chinese, or Simplified Chinese text for free from a JPG, PNG, WebP, or screenshot in your browser.
The image and extracted text stay in this browser. The OCR engine and selected public language model load only after you start.
Drop one JPG, PNG, or WebP image here
You can also paste an image from the clipboard while this tool is open.
One image, up to 10 MB, 12 megapixels, and 8,192 px on either side.
Still JPG/JPEG, PNG, and WebP only. Animated images and files whose extension or MIME type conflicts with their contents are rejected.
Keep this tab open and the screen awake. A mobile browser may pause recognition when it is backgrounded or locked.
This image-to-text tool runs OCR in your browser after you choose a language and start recognition. It is useful for printed notices, clear screenshots, scanned pages, labels, and other horizontal text. The result is plain, editable text, so you can correct OCR mistakes before copying it or downloading a UTF-8 TXT file.
How to use this tool
- Choose, drop, or paste one JPG, PNG, or WebP image.
- Select English, Traditional Chinese + English, or Simplified Chinese + English.
- Rotate the preview until the text is upright.
- Select Extract text and keep the tab open while the OCR engine runs.
- Review and edit the result, then copy it or download a TXT file.
Ways to use it
- Screenshot to text
- Turn text in an app, website, message, or error screenshot into editable plain text.
- Printed notice or scanned page
- Recover horizontal printed text from a clear photograph or a flat page scan.
- Chinese OCR
- Choose the matching Traditional or Simplified Chinese mode for Chinese text that may also contain English and numbers.
- Reuse details accurately
- Extract addresses, reference numbers, email addresses, and URLs, then verify them against the source image.
Supported files and limits
- One still JPG/JPEG, PNG, or WebP file; maximum 10 MB, 12 megapixels, and 8,192 px on either side.
- Large accepted images are reduced to at most 8 megapixels and 4,096 px on either side for browser OCR.
- Printed horizontal text works best. Handwriting, vertical Chinese, tables, columns, exact layout, and historical characters are not supported.
- OCR can confuse similar characters and punctuation. Always compare important names, numbers, email addresses, and URLs with the image.
- The result is capped at 1 MB of UTF-8 text, and a recognition attempt stops after 120 seconds.
Common questions
Which image formats can I convert to text?
The tool accepts one still JPG/JPEG, PNG, or WebP image. It rejects GIF, animated PNG/WebP, HEIC, AVIF, TIFF, BMP, SVG, and PDF.
Can it extract Traditional Chinese text?
Yes. Choose Traditional Chinese + English. That mode is intended for horizontal printed Traditional Chinese and can also read English words and numbers in the same image.
Can it extract Simplified Chinese text?
Yes. Choose Simplified Chinese + English. Traditional and Simplified Chinese use separate recognition modes; the tool does not detect or translate between them.
Can I use a screenshot?
Yes. Clear screenshots with upright, high-contrast text often work well. Small or compressed text may need a higher-resolution capture.
Why are some characters wrong?
Blur, glare, shadows, low contrast, unusual fonts, small type, mixed directions, and compression artifacts can reduce OCR accuracy. Correct the editable result while comparing it with the source.
Does it support handwriting, tables, or exact formatting?
No. This version returns plain text from horizontal printed content. It does not rebuild columns, forms, tables, fonts, or page layout.
Can I extract text from a PDF or HEIC photo?
Convert a PDF page with PDF to JPG/PNG first. For a HEIC photo, create a JPG or PNG copy before using the resulting still image here.