🚀 Ultra Ren'Py Translator: Massive Update (v0.12.5 to v19.2.0)
This is one of the biggest updates in the history of the project! We have completely rewritten major parts of the core engine, introduced real-time translations, integrated advanced AI Fallbacks, and boosted performance across the board.
Here is a summary of all the new features and improvements you can enjoy in this version:
🔴 LIVE Translator (Real-Time Translation)
Play and Translate on the Fly: Introduced a brand-new LIVE Translator system that works similarly to XUnity AutoTranslator. It captures dynamic text, variables, and pop-ups while you play the game and translates them in the background.
Zero-Delay & Caching: We optimized the real-time hook so the game doesn't stutter while translating. Once a string is translated, it's cached instantly for your next playthrough.
Pre-warming System: For larger games, the LIVE Translator pre-loads translations 100x faster, ensuring smooth gameplay without missing any text.
🤖 Smart LLM Fallback (Claude, OpenAI, OpenRouter)
The "Last Mile" Translation: We added a powerful AI Fallback system that uses LLMs (Large Language Models) to translate hard-to-crack slang, idioms, and easter eggs that standard translation engines fail to catch.
Free LLM Options: Integrated OpenRouter, allowing you to choose from 8 high-quality, completely FREE models (like DeepSeek, Llama 3, Mistral) without needing paid API credits.
Separate API Keys: You can now save API keys for Claude, OpenAI, and OpenRouter independently. The UI dynamically adapts and safely stores your keys between sessions.
🧬 Brutal AST Walker (Deep Extraction)
100% Coverage: Replaced simple regex text searching with a deep AST (Abstract Syntax Tree) Walker. It reads the game's compiled code node-by-node, extracting hidden UI elements, conditionals, menu choices, and embedded Python code that were previously impossible to capture.
Autonomous Auto-Fix Loop: The program now analyzes the translation results and automatically re-translates problematic strings using scaled strategies (escalating from Bing to Google Turbo to LLMs) until it hits maximum coverage.
âš¡ Extreme Performance & Optimizations
10x Faster Writing: Implemented buffered writing for translated files. Massive games with over 45,000 dialogue lines that used to take 30 seconds to save now save in just ~3 seconds.
Huge Game Support: The in-game activator has been heavily optimized for games with massive dictionaries. There is no longer any lag or menu stuttering when opening large visual novels.
Fast Startup: We broke down the massive code files into smaller, modular pieces (ui/ panels, dedicated translators, lazy imports). The program now launches 43% faster and consumes less memory on startup.
🎨 UI Overhaul & Quality of Life
Redesigned Home Panel: The main screen has been completely cleaned up. Advanced features (LIVE translation, LLM Fallbacks, Native forcing) are now neatly organized into their own sections with helpful tooltips.
Fullscreen by Default: The app now opens maximized to give you a better view of all the tools and panels.
Better Bug Tracking: We added a structured, readable update log directly into the sidebar so you can see all the new features without leaving the app.
Thank you to all Patreon supporters for making this massive architectural leap possible!