V-Proofix v1.1 release

V-Proofix Changelogs

Version 1.1.0

New Features

• Global Input Blocker: Implemented a comprehensive Keyboard and Mouse hook system (Low-Level API) that temporarily freezes physical input while the application is processing grammar fixes. This prevents accidental tab switches (Alt+Tab) or mouse clicks that could cause the target text box to lose focus.
• Human-Readable AI Models: The application now seamlessly translates backend AI model identifiers into clean, professional display names (e.g., gemma-3-27b-it is shown as Gemma 3 27B, gemini-2.5-flash as Gemini 2.5 Flash) for better aesthetics during API calls.
• Smart Limitation System: Added an automatic fail-safe that rejects texts exceeding 600 words, protecting your API key from hitting rate limits due to excessive payload sizes on single requests.
• Full Localization (i18n): Completely translated the application's user interface. You can now toggle between English and Vietnamese seamlessly in the Settings window without having to restart the application.

UI/UX Improvements

• MacOS-Inspired Indicator Redesign: Expanded the Progress UI window and added a sleek, modern, light-blue glow (DropShadowEffect) around vector icons (Spinner, Checkmark, Error Cross) and primary status text.
• Progress Simulation Illusion: Revamped the visual feedback loop. The application now only shows the boring "Calling API..." state for exactly 1 second. The remaining network latency is masked behind a high-speed percentage simulation (Analyzing... -> Fixing X characters (Y%)...), giving users the illusion of lightning-fast local processing.
• Settings Dashboard Tuning: Fixed window scaling issues. The settings dashboard height is now perfectly proportioned, and user resizing has been securely locked out to maintain strict layout integrity.
• Simplified Messaging: Removed redundant warning text ("Please do not switch tabs") since the new physical InputBlocker makes it completely unnecessary.

Under the Hood

• New Default Brain: Switched the default processing engine to Google's highly impressive Gemma 3 27B (Google's latest 3rd Generation Open Model) instead of Gemini Flash for optimal offline-grade intelligence.
• Robust Model Fallback Chain: Upgraded the backend logic to gracefully degrade. If the Gemma model fails or exhausts its quota, the application will forcefully fall back through a long chain of Gemini models (Gemini 2.5 Flash, Gemini 2.5 Pro, Gemini 2.0 Flash, etc.) ensuring a near 100% processing success rate.
• Rate Limit Cleanups: Stripped out legacy, manual Requests-Per-Minute and Requests-Per-Day tracking logic due to Google API platform limitations, leading to slightly faster initialization times.