Posted November 17, 2025 by Thaluna.Dev
#ai-translator #anime-translator #game-translator #language-learning #Manga #ocr #offline-translator #real-time-translator #translation-tool
This hotfix improves overall stability, OCR accuracy, and UI consistency. It addresses several issues reported by users, especially those related to OCR/translation synchronization.
Some users previously experienced βgarbageβ output or unpredictable behavior when switching between translation models and OCR languages. This was caused by the app desynchronizing the selected translation model and the active OCR engine.
Thaluna now uses a hybrid OCR management system:
β’ When selecting a model for languages with unique alphabets (Japanese, Chinese, Korean), Thaluna will automatically set the correct OCR language to ensure accurate text recognition.
β’ For other translation models (English β Polish, Ollama models, etc.), the app will respect your manual OCR selection for full flexibility.
Fixed an issue where Thaluna could start with the wrong OCR language even if the configuration file specified the correct one.
OCR language is now loaded and initialized properly at startup.
Resolved a layout issue where the translation/OCR buttons could overlap, misalign, or stretch incorrectly when resizing the window.
These elements now maintain a clean, stable layout under all window sizes.
Several internal adjustments were made to improve consistency and reduce random behavior, resulting in a smoother and more predictable translation experience.
Thank you for your continued support, feedback, and patience! π