Loading the interactive tool… It runs in your browser — if it doesn't appear, enable JavaScript.
Pull the text out of a photo of a document, a screenshot of an error message, a whiteboard picture or a scanned page — paste-ready in seconds. Recognition runs on Tesseract, the open-source OCR engine, on our server; images are processed in memory and deleted immediately.
Clear, straight-on shots of printed text give the best results. Tips that genuinely help: good lighting, fill the frame with the text, and avoid skewed angles.
Frequently asked questions
Does it read handwriting?
Barely — Tesseract is built for printed text. Neat block capitals sometimes work; cursive rarely does. For printed documents, screenshots and signs, accuracy is high.
Which languages are supported?
English, Spanish, French and German currently — pick the document's language before extracting, since the language model meaningfully affects accuracy. More languages are easy to add; tell us what you need.
Why is my screenshot text coming out garbled?
Very small text is the usual cause — OCR wants characters at least ~20 pixels tall. Zoom the content before screenshotting, or crop tightly to the text region so the characters are larger relative to the image.