Games built across the AI Playground — some you play with your bare hands and voice, some run agents against each other, some teach a neural network to play for you.
Play in the browser, watch agents compete, or train a network to play for you.
Played with your hands, face and voice, tracked live from the webcam and microphone straight in the browser.
Games where LLM agents compete, deduce, bid and read your mind.
Runners, racers, shooters and brawlers — mostly single HTML files with zero build step.
Memory, real-time strategy, mazes, board classics and card collecting.
Agents that train themselves to play classic games from raw pixels and rewards.
Small single-purpose game POCs — quick Rock-Paper-Scissors builds across different stacks.