OCR It

web edition
engine: idle 100% local · nothing leaves this machine
Source

Transcribe the unselectable

A scanned book, a slide deck, an image-based PDF — anything you can see but can't select. Pin a capture region once, then capture one page at a time; every capture is OCR'd locally and appended to a running transcript you can copy or download.

You can also drop a PDF or images here, or paste a screenshot straight from the clipboard.

How it works
  1. Load a document into the viewer (or share another window).
  2. Drag a rectangle over the text once — refine it with the handles or arrow keys, Enter to pin it.
  3. Press Capture page (or C) once per page — or start an auto-run and let it turn pages itself.
  4. Copy all or download the transcript as .txt.

OCR is performed by a bundled Tesseract build running in this tab. No network, no API key, no upload — ever.

no document
no region pinned
Pin a region first, then capture once per page.
View more demos Get $10 off Kimi K3