Text Grab is a Windows desktop application for OCR text capture and editing. It provides multiple capture modes: full-screen grab, grab frame (interactive region selection), edit text window (with text/spreadsheet/markdown modes), and quick lookup. It integrates multiple OCR engines (Windows built-in OCR, Tesseract, Windows AI) and UI Automation (accessibility tree) for text extraction. The architecture is a monolith with a WPF UI, services layer, utilities, and models. Key entry points include App.xaml.cs, FullscreenGrab, GrabFrame, EditTextWindow, QuickSimpleLookup, and SettingsWindow. Data flows: screen capture or file input → OCR processing → word border rendering → text output with post-processing (templates, patterns, calculations). External dependencies include Windows SDK, Tesseract, ZXing.Net, Magick.NET, PdfPig, Markdig, NCalcAsync, UnitsNet, and others. The codebase demonstrates high craftsmanship with reusable components like the undo/redo system, calculation service, image change detector, automation profile, and custom UI controls. It has extensive test coverage (unit tests and UI automation tests) and CI/CD pipelines. However, some classes are large and tightly coupled, and the Windows-specific dependencies limit portability.