OCR that runs on your machine, not a stranger's server
Most “free OCR” websites upload your document to their server — a strange place for a payslip, a letter or an ID scan to visit. This tool runs the Tesseract OCR engine entirely in your browser (WebAssembly): the first time you extract text, the engine and English language data download once from this website (about 15 MB in total) and are cached; after that, recognition happens fully on your device and your image never goes anywhere.
How to use it
- Add a screenshot, scan or photo with printed text.
- Press Extract text and watch the progress bar — the first run includes the one-time engine download, later runs are far quicker.
- Fix any misreads right in the result box, then copy or download the .txt.
Getting good results
OCR loves contrast and straight lines: a sharp screenshot beats a tilted photo, dark text on white beats white on colour, and 200+ DPI scans beat thumbnails. Handwriting is the honest exception — Tesseract reads print, and reads cursive about as well as a stranger reads your doctor's notes. English only, for now: the language data is what costs the megabytes, and one language done properly beats ten done badly.