Awesome AI Agents › Generative Agents Simulation
WorldSeed is an MIT-licensed multi-agent engine built around emergence rather than fixed workflows. Instead of wiring a directed pipeline of steps, the developer defines a world: the roles agents play, the rules they operate under, what private information each holds, which actions are available, and what consequences follow. Agents then interact inside that world until useful artifacts appear. The project summarizes the model as rules plus differentiated agents plus consequences producing emergence, and the engine itself is scene-agnostic, so the same runtime drives production rooms, simulations, games and fictional settings. The repository documents several demonstration scenes. In Autoresearch, a cohort of specialist agents pursues a rough research idea, proposing hypotheses, running experiments, and peer-reviewing and citing each other in the manner of a research community; the published run targeted validation loss on a small GPT trained on TinyStories and reports 100 hypotheses, 86 experiments and 72 peer-reviewed papers over roughly eleven hours. Every artifact in that run is auditable end to end, linking hypothesis, commit, experiment, verified result, citations and reviewer reasoning into a citation graph. A second scene has one agent study a new API while builder agents produce competing demos, critics reject generic work, audience agents judge usefulness and a curator ships the strongest result together with its trail of attempts and revisions. A third scene is a social simulation about AI-driven layoffs. The maintainers highlight emergent behaviors they did not configure, such as a data specialist drifting into teammates territory once her own lane stopped yielding results. Observation is interactive: a user can watch from above, intervene in the running world, or step into a character. The codebase is primarily Python with a TypeScript front end, and there is a hosted demo. It suits researchers and developers experimenting with agent-based simulation and emergent multi-agent behavior.
https://github.com/AIScientists-Dev/WorldSeed
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.
WorldX turns a single sentence into a running simulated world, generating maps and characters with LLMs and letting autonomous agents act, remember and form stories inside it.
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.