OCR

Read text, QR codes and barcodes out of an image or PDF. Runs in your browser by default. No account, no upload, nothing stored.

Drop an image or PDF here, paste from the clipboard, or

Show word boxes

Notes

Storage: None. On device mode never sends the image anywhere; Precise mode holds it in memory for one request and writes nothing to disk. Reloading the tab loses everything.
On device: Tesseract in a WebAssembly worker. Each page is preprocessed four ways — enhanced, binarised, adaptive and untouched — read in parallel, and the most confident result wins.
Precise: Sends one page to a vision model. Slower and it leaves your device, but it handles handwriting, tables and bad photos that Tesseract gives up on.
Codes: QR, Data Matrix, Aztec, PDF417, EAN, UPC, Code 39/93/128, Codabar and ITF, from a file or live from the camera.
Found: Emails, links, phone numbers, dates, amounts, IBANs, card numbers (Luhn-checked), IPs, tracking numbers, VINs, hashes and passport MRZ lines, pulled out of whatever was read.
Exports: Plain text, JSON, LaTeX source, and a searchable PDF — the original page with an invisible text layer behind it, so it selects and highlights in Preview, Acrobat or any other viewer.
PDFs: Up to 20 pages. A PDF that already has a text layer is read exactly, with no OCR pass at all.