Image to Text (OCR)

Extract text from an image or scan.

Drop a photo, screenshot, or scanned page and pull out the text — then copy it or download a .txt. Recognized locally with PaddleOCR (WebAssembly) in 17 languages, from Korean, Japanese and Chinese to Arabic and Thai; the only network use is a one-time model download per language that's then cached. Your image never leaves this device.

Processes your input on your device without uploading it. Libraries or models may need downloading. No site account is required.