pokemon-player

Automating Pokemon gameplay involves controlling a headless emulator and reading RAM state to guide AI decisions. This Skill uses the pokemon-agent package to run classic ROMs without a GUI, enabling automated play, state checks, and save or load management.

RAM-driven AI play reads game memory state directly from the emulator to inform automated decisions. Combined with optional vision assistance, it allows an AI agent to observe the game environment and execute actions like navigating menus or battling.

Yes, Python 3.10+ is required to run the pokemon-agent package. You must create and activate a virtual environment, install the necessary dependencies including pyboy, and place a valid ROM file in the roms directory before starting the server.

Setting up the server requires cloning the Pokemon agent repository, configuring a Python 3.10+ environment, placing a ROM in the roms directory, and starting the server. It exposes endpoints for state checks, screenshots, saving, and loading progress.

Yes, vision-assisted decision making is supported alongside RAM-state driven observation. While the headless emulator runs without a GUI, the system can process screenshots to optionally guide AI decisions during gameplay.

Automating Pokemon ROMs requires a valid ROM file and a correctly configured Python environment with pyboy. It is limited to classic Pokemon games supported by the emulator and depends on server endpoints for state checks and progress saving.