OASIS is an open-source social media simulator designed for researchers and developers to study complex social dynamics at scale. It allows for the simulation of up to one million agents interacting on platforms like Twitter and Reddit, mimicking real-world user behavior and content discovery.

How It Works

OASIS leverages large language model (LLM) agents to realistically portray user actions, supporting 23 distinct actions such as liking, commenting, and following. It incorporates integrated recommendation systems (interest-based and hot-score-based) to simulate content discovery and platform dynamics. The simulator is built for scalability, adapting to real-time changes in social networks and content for authentic simulation experiences.

Quick Start & Requirements

• Install via pip: pip install camel-oasis
• Requires an OpenAI API key.
• Example agent profile file (user_data_36.json) is available for download.
• Documentation and tutorials are available for detailed instructions.

Highlighted Details

• Supports simulations of up to one million agents.
• Features 23 diverse agent actions for rich interactions.
• Integrates interest-based and hot-score-based recommendation systems.
• Achieved 13x speedup with multi-threading support for LLM inference.

Maintenance & Community

The project is actively developed with recent updates including PyPI release, documentation, and multi-threading support. Community engagement is encouraged via Discord and WeChat.

Licensing & Compatibility

The source code is licensed under Apache 2.0, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

The project is relatively new, with initial release in late 2024. While designed for large-scale simulations, performance and resource requirements for one million agents are not explicitly detailed.