Choose a JPEG, PNG, or WebP image containing printed English text.
Extracted text
How it works
A locally vendored Tesseract.js worker loads a same-origin WebAssembly core and English language model. The source image is visually oriented and bounded before recognition. Progress is shown while the reusable worker processes the page in browser memory.
Limits and assumptions
OCR is an estimate and can make mistakes. Clean, high-contrast printed English works best. Handwriting, blur, decorative fonts, low contrast, rotation, complex layouts, and non-English text may reduce accuracy. Input is limited to 15 MB and 20 megapixels.