Image to Text: Extract Text From Images With OCR
Extract text from photos, screenshots and scanned images (JPG, PNG, WebP) with OCR in your browser, then check it and copy it or save it as a .txt file.
Drop images here, choose them or paste an image
JPG, PNG, WebP, GIF (first frame), BMP and, where your browser can open it, AVIF. You can also paste a screenshot with Ctrl+V (⌘+V on a Mac).
Choose the language, then add images.
Images and recognised text
Notes
- The first time you use a language, your browser downloads the OCR engine (Tesseract.js) and that language's data from leedybox.com: several megabytes, depending on the languages. Tesseract.js keeps the language data in your browser's storage so that it does not need to be downloaded again.
- Your images are processed in your browser and are not uploaded to LeedyBox or anywhere else.
- Accuracy depends on the image quality, the language and the font. Sharp, straight, high-contrast text works best. Handwriting, decorative fonts, very small text and photos taken at an angle are often recognised poorly.
- Only plain text is extracted. Layout, columns, tables, fonts and formatting are not kept, and the reading order of complex pages can be wrong.
- The confidence value is the OCR engine's own estimate, averaged over the words. It is not a measure of how many errors the text contains.
- Images are read one at a time. Images larger than about 25 megapixels are scaled down before reading to limit memory use, and very large images can still fail on phones.
- For PDF files, use PDF to Text: it reads the text layer of a PDF and uses OCR only for pages that do not have one. Open PDF to Text
How to use
- Choose the language of the text and, if the images also contain a second language, choose it as well.
- Add one or more images: drop them on the box, choose them, or paste a screenshot with Ctrl+V (⌘+V on a Mac) or the "Paste image" button.
- Wait while the images are read one at a time. The first time, the OCR engine and the language data are downloaded, which can take a moment.
- Check the recognised text against the image and correct it in the text box where needed.
- Copy the text or download it as a .txt file. With several images, you can copy or download all texts together.
Frequently asked questions
Are my images uploaded?
No. The images are processed in your browser and are not uploaded to LeedyBox. Only the OCR engine and the language data are downloaded from leedybox.com, and those downloads do not contain your images.
Why does the recognised text contain mistakes?
OCR estimates which characters an image shows. Blurry or small text, low contrast, skewed photos, unusual fonts and the wrong language setting all cause errors. Try a sharper image, crop it to the text, choose the right language or turn on the image options, and check the result.
Can it read handwriting?
Only to a limited extent. The engine is trained on printed text, so handwriting is usually recognised poorly. Neat, separated block letters work better than joined-up writing.
Is there a limit on the number or size of images?
There is no fixed number, but the images are read one after another and each one takes a few seconds or more. Images larger than about 25 megapixels are scaled down first, and very large images can fail on devices with little memory.
Does it keep the layout, tables or formatting?
No. The tool extracts plain text only. You can keep the line breaks as they appear in the image or join lines into paragraphs, but columns, tables and formatting have to be rebuilt by hand.