How to Extract Text from an Image (Free OCR in Your Browser)
A photo of a printed letter, a screenshot of an error message, a receipt you need for your expenses: the text is right there, but you can’t select it. Typing it out again takes time and adds mistakes. OCR does it for you in a few seconds.
This guide shows how to extract text from any picture with the free Image to Text converter, which settings to choose, and how to get an accurate result.
What is OCR?
OCR stands for optical character recognition: software that looks at the shapes in a picture and recognises them as letters, numbers and words. Modern OCR reads whole lines at once instead of one letter at a time, so it copes well with ordinary photos and screenshots.
Our tool uses Tesseract, the well-known open-source OCR engine, running entirely inside your browser. Your images are never uploaded anywhere.
How to extract text from an image
- Open Image to Text and add your picture: drop it on the box, click Choose images, or copy it and press Ctrl+V (Cmd+V on a Mac). Screenshots can be pasted straight in. JPG, PNG, WebP, GIF, BMP and HEIC (iPhone photos) all work, up to 20 images at a time.
- Choose the Text language. It matters: the reader uses the language’s alphabet and common words to decide what it sees.
- Click Extract text. The first time you use a language, your browser downloads its data (1 to 3 MB) and keeps it for next time.
- The text appears next to the picture, with each line it found boxed in blue. Correct anything if needed, then click Copy text or Download .txt. With several images, Copy all and Download all (.txt) give you every text in one go.
Which languages can it read?
English, Spanish, German, French, Urdu, Hindi, Arabic, Chinese (Simplified), Japanese and Korean. Urdu and Arabic text is shown right to left.
Many screenshots mix two languages, for example Hindi with English brand names or Arabic with English numbers and web addresses. When you pick a language other than English, tick Also read English and both are read together. It is ticked for you when you choose a language with another alphabet.
The layout setting
- Automatic finds the blocks, columns and lines on its own. Best for documents, letters, articles and most screenshots.
- One block of text reads the picture as a single paragraph. Try it for a cropped paragraph or a short quote if Automatic splits lines in odd places.
- Scattered text (signs, labels) looks for words anywhere, without expecting neat lines. Use it for street signs, product labels, posters and pictures with text in different spots.
Changed your mind? Pick another language or layout and click Read again. Reading the same image again is free and doesn’t count towards your daily limit.
How to check the result
Above the text you see the number of words and a Confidence score: how sure the reader is overall. Above 90 % usually means the text is almost perfect; below 60 % means the picture was hard to read.
Words the reader wasn’t sure about are marked in amber on the picture. Look at those first: they are where mistakes hide, such as a 0 read as O or 1 as l. You can correct the text right in the box, and Undo my changes brings back what was read.
Tips for better results
- Sharp and well lit. Blur and shadows are the main causes of mistakes. Hold the phone steady and avoid glare on glossy paper.
- Straight on. Take the photo from directly above the page, not at an angle.
- Crop to the text. Pictures, logos and backgrounds around the text can be mistaken for letters.
- Big enough. Text should be clearly readable when you look at the image yourself. Small screenshots are enlarged automatically before reading.
- Printed text works best. Handwriting is read poorly, especially joined-up writing.
Is my image uploaded?
No. The reader runs in your browser, so your pictures and their text never leave your device. That makes it safe for receipts, letters, ID numbers and anything else private.
Frequently asked questions
Can it read a scanned PDF?
Not directly: it reads images. Take a screenshot of each page (or save the pages as images) and add those. To check two versions of a PDF that already has text, use Compare PDF Files.
Does it keep the formatting?
It keeps lines and paragraphs, and columns of numbers on receipts usually stay lined up. Fonts, colours and bold text are not kept: you get plain text, ready to paste anywhere.
Why did it find no text?
The picture may be too blurry or small, or the wrong language may be chosen. Try a sharper picture, check the language, or switch the layout to Scattered text. Images where no text is found are never counted.