Home › Image Tools

Image to Text (OCR)

The first time, language data (about 3 MB for English) is downloaded once from jsDelivr (cdn.jsdelivr.net). Your images are never sent.

Files are processed in your browser and never uploaded
Drop images with text here or click to choose (multiple allowed)JPG · PNG · WEBP · HEIC · you can also paste screenshots (Ctrl+V)
Lines
The recognition engine (Tesseract.js) loads from this site and the language data from the jsDelivr CDN (cdn.jsdelivr.net/npm/@tesseract.js-data). It's cached in your browser, so next time it starts right away. Your images are never sent anywhere.

Text out of pictures

Pull the text out of a printed notice, a photo of lecture slides, a screenshot or a scanned document without retyping it. This tool runs the open-source Tesseract engine (Tesseract.js) inside your browser. Images never go to a server, so it's safe for personal documents.

Language data is downloaded once

Recognition needs trained data for each language. The first time, the English data (about 3 MB; Korean is about 1.6 MB) is downloaded from the jsDelivr CDN and cached in your browser, so later runs start immediately. Only the language data travels — your images are never sent.

Tips for better accuracy

Cleaning up the result

“Join line breaks within paragraphs” reconnects sentences split by line breaks in the image. Edit the result right in the box, then copy it or download it as TXT. A low confidence score means more mistakes, so proofread. Handwriting and vertical text are not recognized well.

Sources

Frequently asked questions

The first run takes a while.

The language data is downloading. It's cached in your browser, so later runs are faster.

Does it work offline?

Once the language data is cached, usually yes. The first run needs an internet connection.

Can it read PDFs?

Convert PDF pages to images with PDF to JPG first, then drop them here. If the PDF has real text, Extract Text from PDF is faster and exact.

Related tools