GregoryCannon/StackRabbit

An AI for playing NES Tetris at a high level. Based primarily on search & heuristic, with high quality board evaluation through value iteration.

What it solves

StackRabbit is an AI designed to play NES Tetris at a high level of skill, automating the game's decision-making process to achieve high scores.

How it works

The system uses a combination of search and heuristics, utilizing value iteration to perform high-quality board evaluation. The architecture consists of a Node.js server that handles request parsing and delegates tasks to C++ modules, which perform the core AI computations at significantly higher speeds than a JavaScript implementation.

Who it’s for

It is intended for players and developers interested in high-level NES Tetris AI, as well as those who want to experiment with AI-driven gameplay on emulators or real hardware.

Highlights

  • Multi-platform support: Includes clients for the FCEUX emulator (Lua) and real NES consoles via Raspberry Pi (Python).
  • High-performance computation: Core AI logic is implemented in C++ for maximum speed.
  • Web accessibility: A separate web client allows users to to draw boards and query the AI for optimal placements.

Related

  • Project
  • Project
  • Project
  • Project
  • Project