Show HN: Watch a neural net learn Clash Royale defense in the browser

Category: ai-ml

Tags: reinforcement-learning, game-engine, simulation, cpp, python, webassembly

Score: 7.0/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 5)

ClashRoyaleEnv is a high-performance, deterministic Clash Royale battle simulator built in C++ with Python bindings, designed for reinforcement learning research. It includes a full RL training stack, lookahead search, a computer vision bridge for screen capture, and an interactive WebAssembly demo called Reflex Lab where users can watch a small neural network learn to defend in the browser. The project stands out for its simulation speed (20,000× real time) and comprehensive tooling, making it a compelling platform for game AI experimentation.

Target audience: AI researchers, reinforcement learning practitioners, game AI hobbyists, and developers interested in simulation environments

Repository: https://itzik123.github.io/ClashRoyaleAi/lab/ · Python · MIT · 24 stars

View on Hacker News