🤖 Learning Agent Gridworld: 2D Q-learning

A 2D gridworld where a reinforcement-learning agent learns by trial and error with tabular Q-learning. Paint walls and pits, drag the goal and start tiles, and watch the value map, policy arrows and learning curve re-adapt when the world changes.

2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install