joonspk-research/generative_agents

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

21995 stars · 3076 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

The official code release for the Stanford 'Generative Agents' research paper, implementing computational agents powered by LLMs that simulate believable human behavior in a Smallville game environment. It consists of a Django-based environment server and an agent simulation backend that requires an OpenAI API key.

Use cases

• simulate believable human-like agents in a virtual town
• reproduce the generative agents research paper experiments
• build LLM-powered agent simulations with memory and planning
• run the Smallville demo animation of agent behavior
• study social simulation with large language models
• prototype agent architectures like memory streams and reflection

When to choose

• you want to replicate or extend the canonical generative agents research
• you need a reference implementation of LLM agents with memory, planning, and reflection
• you are doing academic work on believable agent simulation

When to avoid

• you need a production-ready or actively maintained agent framework
• you want to avoid OpenAI API costs or vendor lock-in
• you need a lightweight library rather than a full two-server simulation setup

Facets

application · maturity maintenance

agent-framework llm-inference simulation game-engine artificial-intelligence large-language-models simulation python cross-platform generative-agents llm-agents social-simulation smallville research-paper believable-agents ai-agents research web-server

Member repositories

For agents

markdown · JSON · MCP: product_card(name="joonspk-research/generative_agents")