Awesome AI Agents › Generative Agents Simulation
WorldX is an alpha-stage TypeScript project that converts one sentence of description into a simulated world rendered as a pixel game with React 19 and Phaser 3 on Node.js 18 or newer. It uses four separately configured model roles: an orchestrator that designs world structure, characters and rules; an image generation model that draws map art and character portraits; a vision model that reviews map quality and locates regions and elements; and a cheaper simulation model that drives character behaviour at runtime. Each role takes its own base URL, API key and model id from environment variables, and every role except image generation speaks the OpenAI-compatible chat/completions protocol, so providers such as OpenRouter, Google AI Studio or DeepSeek can be mixed freely; the image role additionally supports a google-native provider mode. Once a world exists, characters act on their own: they make decisions, interact with scenes, build relationships, hold conversations, keep memories of past events and develop behaviour patterns from those memories, so narratives emerge without a written script. Simulation continues across day and night cycles over multiple in-world days, and a single world can carry several parallel timelines. The user can intervene at any moment in a god mode by broadcasting events, editing a character persona or memory, or holding a direct conversation with any character, then observing how the world reacts. Two pre-generated worlds ship with the repository so the simulation can be run after configuring only the simulation model, while full creation happens through a web interface on localhost:3200 or a command-line create script. Interface and generated content are bilingual in Chinese and English. The project targets people experimenting with emergent narrative, generative agent simulation and procedural world building.
https://github.com/YGYOOO/WorldX
This project provides a simulation framework for generative agents that mimic believable human behaviors within an interactive game environment, enabling running, saving, replaying, and demonstrating agent-based simulations.
Voyager is an open-ended embodied agent powered by large language models that autonomously explores and learns diverse skills in Minecraft through lifelong learning and an evolving skill library.
SFighterAI is a deep reinforcement learning-based AI agent designed to master and beat the final boss in Street Fighter II: Special Champion Edition using only game screen pixel data.
A comprehensive Godot framework for creating smooth and complex AI motion using steering behaviors in both 2D and 3D games.
Control room for LLM agent societies: pause and scrub a running simulation, question individual residents, inject instructions into the next step, and export replayable experiment packs.
PyGame Learning Environment (PLE) is a Python-based reinforcement learning platform that provides a variety of games and tools to facilitate the development and testing of reinforcement learning algorithms.
A locally runnable, low-cost implementation of generative agents simulating human-like behavior in an interactive Dungeons & Dragons setting using large language models.
GPTRPG is a proof-of-concept project featuring a GPT-based AI agent that autonomously interacts within a simple RPG-like environment using the OpenAI API.