Image to Text (OCR)

Turn photos, screenshots and scans into editable, copyable text - in 16 languages. Recognition happens in your browser, so sensitive documents stay private.

Drop an image here or click to browse

Screenshots, receipts, book pages, whiteboards, signs
Extracted text

First run loads the language model (a few MB, cached afterwards). More languages = slower, so tick only what you need.

How it works

  1. Drop an image

    Any photo or screenshot with printed text.

  2. Choose languages

    English is preselected; add French, Arabic, Japanese and 13 more as needed.

  3. Copy the text

    Recognition runs on-device with progress shown, then copy or download the result.

FAQ

How accurate is it?
For clean, straight, well-lit printed text it's very good - this is Tesseract, the open-source OCR engine used in industry, compiled to run in your browser. Handwriting, heavy blur and skewed low-light photos will have errors; crop tightly and increase contrast for best results.
Does it keep tables and layout?
It returns plain reading-order text, not tables or columns as structure. For layout-faithful copies of PDFs, use PDF to JPG or PDF to PPT instead.
Is my image uploaded?
No - the recognition engine and language data run locally. Only the models are downloaded from a CDN; your image never travels.