pokemon-player

To automate playing Pokemon games using a headless emulator, this Skill extracts RAM state to drive gameplay decisions without a visible UI. You need to clone the pokemon-agent project, set up a Python 3.10+ environment, and provide a ROM file to start the server.

RAM-driven AI for Pokemon emulation is a technique where the agent reads memory addresses directly to understand game state. This removes the need for visual screen parsing, allowing the AI to make decisions and guide gameplay based on exact memory values.

Yes, headless Pokemon automation requires a Python 3.10+ virtual environment. You must create this environment, install the necessary pokemon-agent package dependencies, and place your ROM file in the designated roms directory before starting.

Yes, you can observe an AI agent playing Pokemon in real time using an optional live dashboard. While the emulator runs headlessly, the dashboard provides real-time observation of the gameplay session and health checks for the running server.

To generate training data from Pokemon gameplay, you apply this Skill to automate sessions via a headless emulator and RAM reads. The integrated memory tagging and save/load workflow track progress across sessions to capture the gameplay data.

Using RAM reads for Pokemon AI automation requires specific memory addresses for each game version, meaning the approach is limited to supported ROMs. It also requires a running server with health checks and cannot process visual elements not mapped to memory.