Document OCR
Extract editable text from scanned documents and images.
Find the words. Bring them into focus.
Browser-based recognition · Engine and language files may need downloading. Review every result.
Let the page
speak again.
Recognize printed text in scans and images, compare it with the source, and save the corrections you choose.
How to extract text
- Choose the language. Set the recognition language and optional grayscale/contrast processing.
- Open a scan or image. Load a supported file or paste a screenshot. Engine files may download on first use.
- Read alongside the original. Navigate the page preview and correct the combined text. Check names, numbers and punctuation carefully.
- Take the words with you. Copy the reviewed text or download a plain-text file before closing the page.
What files can I recognize?
Load a PDF, PNG, JPEG, WebP or BMP image, or paste an image from your clipboard. Files must be smaller than 50 MB; PDFs are limited to 20 pages and a combined rendered size of 50 million pixels. Device and individual image limits also apply.
Which languages can I choose?
English, Spanish, French, German, Italian, Portuguese, Simplified Chinese and Japanese are available. Choose the language before loading your file. Recognition quality depends on language, scan clarity and typography.
Can I correct the recognized text?
Yes. Review the editable text beside the source preview and make corrections before copying or downloading it. Multipage output includes page separators; preview navigation does not switch the combined text editor.
Does this create a searchable PDF?
No. The output is editable plain text that you can copy or download as a UTF-8 TXT file. It does not rebuild your document layout or add a searchable text layer to a PDF.
Where does recognition happen?
Document processing and recognition run in browser memory. OCR engine files, PDF libraries and language models may download from external services. Review results before use; the confidence score is an engine estimate, not a guarantee of accuracy.