GenWorlds is an event-based framework for building collaborative, autonomous multi-agent systems. It targets developers and researchers seeking to create interactive AI environments inspired by "Generative Agents," enabling agents to communicate asynchronously, interact with objects, and form memories, ultimately executing complex tasks collectively.

How It Works

GenWorlds utilizes a WebSocket server for asynchronous communication between AI agents and their environment. Agents, powered by Langchain and OpenAI's GPT API, are goal-oriented and can interact with shared objects. The framework incorporates dynamic memory management, allowing agents to learn from past experiences, and is designed for scalability through threading.

Quick Start & Requirements

• Install: pip install genworlds
• Prerequisites: OpenAI API key, Qdrant instance.
• Additional UI: GenWorlds-Community provides a ChatGPT-like UI.
• Docs: GenWorlds Docs

Highlighted Details

• Event-based communication via WebSocket for real-time agent interaction.
• Inspired by "Generative Agents" research and AutoGPT.
• Customizable environments with interactive objects.
• Dynamic memory management for agent learning.

Maintenance & Community

The project is under active development and welcomes contributions. Enterprise support is available through Yeager.ai.

Licensing & Compatibility

Released under the MIT License, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

GenWorlds is in alpha, with potential for breaking changes. Usage can incur significant costs due to OpenAI API calls. Users are solely responsible for managing API costs, ensuring actions align with legal and ethical standards, and assuming all associated risks.