1.0.0177.2 MB
MIT
strict
core22
Offline screen region capture and optical character recognition grabber
Offline Screen Region OCR is an independent open-source Linux alternative to
Capture2Text and Text Grab for instant offline screen-region text extraction
and clipboard capture. It is not affiliated with, sponsored by, or endorsed by
the authors of Capture2Text or Text Grab.
FEATURES
- Translucent Crosshair Region Snipper: Capture any screen rectangle across
single or multi-monitor X11 desktops using a fullscreen rubber-band overlay
or global X11 hotkey (Ctrl+Alt+T), or load PNG/JPG files and clipboard images.
- Image Preprocessing Pipeline: Automatic 1x/2x/3x smooth upscaling for crisp
small UI fonts, grayscale contrast normalization, optional dark-background
auto-inversion, Otsu adaptive thresholding, and quiet border padding.
- Bundled Offline Tesseract 4 OCR Engine: Zero cloud uploads — ships complete
offline English (eng) and Vietnamese (vie) language models with selectable
page segmentation modes (Auto, Single Block, Single Line, Sparse Text).
- Instant Clipboard & Live Confidence Metrics: Copies recognized text directly
to the system clipboard alongside word count, character count, and mean
Tesseract per-word confidence score.
- SQLite OCR History & Export: Browse searchable capture history with saved
region thumbnails, re-copy past extractions, and export to TXT, Markdown,
or UTF-8 CSV files.
SUPPORT & DONATE
- Created by Thanh Nguyen — https://buymeacoffee.com/thanhnv128b
Capture2Text and Text Grab for instant offline screen-region text extraction
and clipboard capture. It is not affiliated with, sponsored by, or endorsed by
the authors of Capture2Text or Text Grab.
FEATURES
- Translucent Crosshair Region Snipper: Capture any screen rectangle across
single or multi-monitor X11 desktops using a fullscreen rubber-band overlay
or global X11 hotkey (Ctrl+Alt+T), or load PNG/JPG files and clipboard images.
- Image Preprocessing Pipeline: Automatic 1x/2x/3x smooth upscaling for crisp
small UI fonts, grayscale contrast normalization, optional dark-background
auto-inversion, Otsu adaptive thresholding, and quiet border padding.
- Bundled Offline Tesseract 4 OCR Engine: Zero cloud uploads — ships complete
offline English (eng) and Vietnamese (vie) language models with selectable
page segmentation modes (Auto, Single Block, Single Line, Sparse Text).
- Instant Clipboard & Live Confidence Metrics: Copies recognized text directly
to the system clipboard alongside word count, character count, and mean
Tesseract per-word confidence score.
- SQLite OCR History & Export: Browse searchable capture history with saved
region thumbnails, re-copy past extractions, and export to TXT, Markdown,
or UTF-8 CSV files.
SUPPORT & DONATE
- Created by Thanh Nguyen — https://buymeacoffee.com/thanhnv128b
Update History
1.0.0 (1)14 Nov 2026, 08:15 UTC
14 Nov 2026, 07:50 UTC
14 Nov 2026, 07:49 UTC
14 Nov 2026, 08:15 UTC


