Scan → choose languages → extract → review
Use Upload files for images or PDFs. On a phone, Scan a page opens your camera. On a laptop, it opens your webcam. Photograph a flat, well-lit page, keep it straight, and crop out the background.
Languages & accuracy
English, Urdu, Sindhi and Arabic models are included. Additional languages download when selected. OCR works best on clear printed text. Urdu Nastaliq, handwriting, stamps, mathematical notation, tables and multiple columns can be unreliable. This app does not translate or invent missing words.
Your documents stay here
Recognition runs in your browser. There is no document-upload server, analytics or account. Optional saved drafts contain your edited text and title in this browser only; they do not sync between devices. Delete a saved draft to erase it. Source images disappear when you close or reload the page.
Network & offline use
Hosting still sees ordinary page requests. Additional language models come from jsDelivr; no document contents are sent. After the app's offline files are ready, it can reopen offline. English, Urdu, Sindhi and Arabic are included in that offline pack. Extra languages need a first online extraction and browser storage may be cleared or evicted.
Install on phone or laptop
Use Install app when supported. On iPhone/iPad, use Safari → Share → Add to Home Screen. On Android or a desktop Chromium browser, use the browser menu → Install app. Always deploy over HTTPS. OCR needs a modern browser with WebAssembly.
Markdown output
Extraction preserves text and paragraphs with optional page headings. Add formatting in the editor. Tables and original document styling are not reconstructed automatically. Download .txt strips common Markdown formatting.
Checking offline availability…