An independent Android app for capturing, editing and collecting text. Accessibility capture and offline OCR work together, with English and French models bundled in the app.
The useful details
Capture visible app text, or use offline OCR for images and PDFs.
Edit a selection, collect across captures, then copy or share.
English and French OCR models are bundled. The app needs no internet permission.
The source is public under Apache-2.0. There is no signed public release linked here. The README explains building it and trying a debug build; a CI artifact is not a stable public download.
Before you use it
Screen capture needs accessibility access and an explicit action from you; accessibility events do not collect text in the background.
Normal text capture skips password fields and their children. Android can refuse screenshots of protected windows; Inkjack does not bypass that protection. OCR cannot identify every sensitive word drawn into an image.
OCR can miss or misread words. Review and edit the result before using it; Rescan with OCR is available when an app exposes only part of its text.
Screenshots, extracted text and collections are temporary, not a saved document library. A collection can disappear when Android ends the process. PDF imports are limited to 20 pages.
Capture targets the default display. There is no automatic scrolling, unattended harvesting, QR recognition or built-in translation.
Offline OCR does not make every destination private. Copy sends text to Android’s clipboard; Share sends it to the app you choose. That app, and your keyboard while editing, have their own privacy settings.