Ai FTB Translator

Version 1.3.0 - The Intelligence & Reliability Update

This is a massive stability and intelligence overhaul. This update focuses on squashing dozens of annoying bugs, making the AI smarter and more reliable, and giving you more control over the translation process. The mod is now more robust and dependable than ever before!

✨ New Features & Major Improvements

• Dynamic Model Selector: You are no longer limited to manually typing a model name! When using a compatible provider (like the default llm7.io), a new button appears next to the "Model Name" field. Clicking it fetches a live, up-to-date list of all available models directly from the API, allowing you to switch between them with a single click.

• Expanded Language Support (16 Languages!): The entire mod interface is now fully translated into 16 languages! All config screen text, tooltips, and pop-up notifications will now appear in your selected language. Support has been added for: Polish, German, Spanish, Brazilian Portuguese, Portuguese, Turkish, Ukrainian, Hindi, Russian, Arabic, French, Simplified Chinese, Japanese, Korean, Italian, and Dutch.

• Added Force Re-translate Functionality: A powerful new feature has been added! If a translation fails or seems incorrect, simply hover over the quest and press the 'R' key (by default). This will instantly clear the cache for the quest's title, subtitles, AND description and request a fresh translation from the AI.

🧠 AI & API Intelligence Upgrades

The core of the mod is now significantly smarter, leading to fewer errors and better translations.

• Advanced Prompt Engineering: The instructions sent to the AI have been completely rewritten to provide game-specific context (e.g., "Minecraft," "FTB Quests"). This drastically reduces translation failures on sensitive words like "Kill" by teaching the AI to understand the game's context.

• A Note on Translation Speed: For improved security and to accommodate the more detailed prompts, the time to receive a translation may be slightly longer than in previous versions. This is a deliberate trade-off for much higher quality and more reliable results.

• Smart API Rate Limiting: The mod now intelligently spaces out API requests to prevent overwhelming free APIs. This virtually eliminates the constant PARSE_ERROR issues that were caused by sending too many requests at once.

• Improved Bad Translation Filter: The filter is now much better at detecting and rejecting "junk" data from the API (like the {"annotations":...} error), preventing it from ever being cached or displayed.

🐛 Bug Fixes & Technical Improvements

• Internal Code Refactoring: The mod's internal file structure has been reorganized for better maintainability and to make future updates easier.

• Double Toast Fix: Fixed the bug that caused the "Config Saved" notification to appear twice when closing the menu.

• Squashed PARSE_ERROR Bugs: The combination of better prompt engineering and rate limiting has fixed the vast majority of cases where the AI would return junk data, leading to a much smoother experience.