Extract printed English or Russian text from PNG, JPG, and WebP images. Copy recognized text from a screenshot or scanned page.
How to use
Choose a clear image, select the recognition language, and start OCR. Compare the result with the image before copying names, codes, or figures.
Example
Read a photographed notice, then check similar characters such as O and 0 before reusing the text.
FAQ
Are images uploaded?
Recognition runs in your browser and the image is not uploaded. Blurry photos, handwriting, and complex columns reduce accuracy.
Related tools
A practical task
Crop to the text region and keep lines as level as possible. A sharp smaller crop often works better than a large photograph with a distant sign. After recognition, compare dates, prices, and character pairs that are easy to confuse.
Another useful question
How do I handle two columns?
Recognize clear regions separately when the reading order becomes mixed. Join the reviewed text afterward rather than accepting interleaved lines as a continuous sentence.
Details for interpreting the result
| Detail | Explanation |
|---|---|
| OCR and text extraction | Text extraction reads text already present in a file. OCR recognizes shapes in an image and guesses characters. Scans and photographs need recognition; ordinary text PDFs usually need extraction instead. |
| Fixed page documents | PDF stores a page layout rather than an editable document model. Visual placement does not always define reading order. Keep an editable source when future changes to wording, headings, or tables are expected. |
| Pixels and text | An image stores pixels, not a searchable text document. A text image can be convenient for display but harder to search, select, translate, or read with assistive tools. Keep a text alternative for important information. |
| Image dimensions | File size and pixel dimensions measure different things. A compressed image can still contain many pixels and need substantial memory when decoded. Inspect both dimensions and file size when preparing image input. |
| Font coverage | A font contains only the glyphs it supports. A Unicode character in the source does not guarantee that every export font can draw it. Inspect accented names, non Latin text, and emoji in the final document. |
| Paragraphs and lines | A manual line break and a paragraph boundary serve different purposes. Export tools can map them differently. Inspect lists, addresses, and intentional blank lines rather than treating every visible break as interchangeable. |
| Source boundaries | A summary or rewrite works from the material you provide. A URL alone does not supply the full article in this interface. Include the relevant passage and keep citations attached to the claims they support. |