Image to Text Converter (OCR)

The words are in the picture. This reads them out.

An image to text converter reads the words inside a photo, scan or screenshot and hands them back as plain text you can copy, edit and search. The reading is done by OCR, optical character recognition, a trained eye that looks at letter shapes and decides what they say. No more retyping a page because it arrived as a picture.

Every day, text gets trapped in images: the photographed notice, the scanned form, the screenshot of an address. Retyping is slow and quietly invents errors, a wrong digit in a phone number, a name spelled almost right. Reading the words out by machine, then checking them, beats typing from scratch.

How to convert an image to text

1. Drop your image, PNG, JPG or WebP, into the box above, or click and pick it.
2. Press Read the Text. The first run also loads the English word data, which can take a moment.
3. Watch the progress as the reader works through the picture.
4. Read the result in the preview box, and copy it straight from there.
5. Download the .txt file if you want the words kept as a file.
6. Check names, numbers and dates against the picture. OCR is good, eyes are the proofreaders.

What makes OCR read well, and what makes it stumble

Sharp, straight and well-lit wins. A flat scan or a clean screenshot reads close to perfectly, because the letters stand still and behave. A tilted phone photo of a curved page is the hardest meal you can serve it: shadows bend the lines, glare eats letters, and small print blurs into guesses. Straighten the shot, fill the frame with the page, keep the light even, and the same tool reads like a different, smarter tool. Handwriting is the known cliff edge: neat printed text reads well, most handwriting reads poorly or not at all, and no honest OCR page should pretend otherwise.

The honest limits of image to text

The output is the words, in reading order, as plain text. Columns can pour into each other, because the reader follows lines, not layouts. Tables come out as lines of words, not as a grid: when the grid itself matters, Image to Excel rebuilds rows and cells instead. The reader here knows English. Text in other scripts needs a reader trained for that script, and mixing scripts in one picture confuses any reader. Unusual fonts, text over busy photos and tiny captions all lower accuracy. Treat the result as a very fast first draft that still wants one human pass.

Jobs people do with image to text

A clerk pulls the words out of a photographed circular so the office can quote it exactly. A student lifts text from a library book photo for notes, with the citation written by hand like a scholar. A driver copies an address off a screenshot instead of squinting and typing it digit by digit into the maps app. A support worker grabs the words from an error screenshot to search for the fix.

Read on your device, sent nowhere

The reading engine runs inside your browser. Your image is never uploaded to a server, an unusual and important point for OCR, since most rival tools read your pictures on their machines. Here, and across the whole converter family, the file stays home.