downthecrop/pokemon-llm

AI Summary: Improve the code quality of a Pokemon game UI and LLM integration project. This involves automating OS detection and mGBA checks in the boot process, improving overall code quality, implementing a configuration file for keys and models, adding a launch argument for AI provider selection, and managing benchmark savestate files more effectively.

AI Summary: This project aims to create a web UI and tools to allow Large Language Models (LLMs) to interact with Pokémon games (Gen 1, 2, and 3) emulated using mGBA. Current functionality is limited to Gen 1. The main tasks include adding support for Gen 2 and 3 (requiring minimap and memory extraction), improving the system prompt for LLMs, optimizing data extraction methods, and enhancing the web UI for better readability and easier game capture integration.