This tool extracts English text from a screenshot, scan or photo using Tesseract OCR in your browser. The image is not uploaded. The first run downloads the OCR engine and the English language model from this site, which is why the first result can take longer than later attempts.

How to extract text

Use the best source you have. A straight, well-lit scan at a reasonable resolution gives better results than a tilted phone photo with a shadow.
Add the image. JPEG, PNG and WebP are supported. Very large images are reduced to a working size to keep the browser responsive.
Start extraction. The progress panel reports loading, initialisation and recognition. Cancel if you selected the wrong file.
Check the result. Compare names, numbers, dates and punctuation with the image. OCR confidence is a signal, not a guarantee. Low confidence often points to a difficult source.
Copy or download. The text can be copied to the clipboard or saved as a UTF-8 text file. The original image is not included in the download.

What OCR can and cannot do

OCR recognises shapes and maps them to characters. It works best with printed text, clear contrast and a normal reading order. It can struggle with handwriting, decorative fonts, very small text, tables, multi-column layouts, stamps, receipts and text over a busy image. It may read “0” as “O”, “1” as “l” or merge two columns into one line.

The tool does not automatically rotate or deskew the page, remove shadows, detect columns or reconstruct a table. If the source is rotated, rotate it in an image editor first. If the page has two columns, crop one column at a time. If the text is faint, increase contrast before uploading. These steps often matter more than choosing a different OCR engine.

Good uses

Copying text from a screenshot when the original is unavailable.
Making a rough transcript of a printed notice or receipt.
Extracting a paragraph from a scanned handout for your own notes.
Checking whether a scan contains readable text before sending it.
Getting a starting point for a document you will proofread.

Limits and accuracy

English only: the bundled model is English. Other languages may produce poor or meaningless output.
First-load cost: the engine and language data are a few megabytes. On a slow or metered connection, wait or use a smaller image.
No handwriting guarantee: cursive and mixed handwriting are especially unreliable.
No layout preservation: paragraphs, tables and columns may be flattened.
No legal or medical advice: never rely on OCR alone for a prescription, contract, identity document or legal filing.

Privacy

The image is processed in this tab. The image bytes are not sent to OmniEmu. The page does load JavaScript, WebAssembly and language files from this site; those requests do not include your image. The extracted text stays in the tab until you reset or close it. Download or copy what you need first. Ordinary hosting logs are separate and are described in the privacy policy.

Frequently asked questions

Why is the first run slow?

The browser has to download the OCR engine, WebAssembly core and English language data, then initialise them. Later runs in the same session can reuse the loaded engine if the page keeps it alive. A slow connection or a low-memory phone can add to the delay.

Why is some text missing?

The source may be rotated, low contrast, too small, handwritten or arranged in columns. Try cropping the text area, increasing contrast, straightening the page or using a larger scan. Do not assume the missing text does not exist.

What does the confidence number mean?

It is Tesseract’s internal confidence estimate. It is not an accuracy percentage and it can be high for wrong text. Use it as a hint to check the result, not as proof that the transcription is correct.

Can it read PDFs?

This tool accepts images only. A scanned PDF must first be converted to page images. A text-based PDF does not need OCR; use a PDF text extractor instead.

Is the text saved?

No. There is no account and no server-side history. Copy, download or paste the text into your own document before closing the tab.