Extracts text from a photo or screenshot using Tesseract's neural-network OCR engine, running entirely in your browser via WebAssembly. Nothing is uploaded.
Drop an image here, or click to choose one
How it works
Text extraction runs on Tesseract.js, a WebAssembly build of the open-source Tesseract OCR engine, entirely inside a Web Worker on your device. The first scan (or any time you switch languages) has to download and initialize that language's trained data model before it can read anything; after that, the same worker is reused for subsequent images in the same language.
FAQ
Why does the first scan take noticeably longer than the next one?
Tesseract.js has to fetch and load the WebAssembly engine plus the selected language's trained model before it can recognize anything — once that's loaded, the same worker handles later scans in that language much faster.
Does the language I pick actually affect accuracy?
Yes — each language loads a separate trained model, so scanning French or Chinese text with English selected will produce garbled output even on a clean, well-lit image, since the model is matching against the wrong character patterns.
Free, unlimited, no account. If this saved you a paid tool, a small thanks is appreciated.