Agentshire  by Agentshire

AI agents visualized as living 3D NPCs in a customizable game town

Created 1 month ago
336 stars

Top 81.9% on SourcePulse

GitHubView on GitHub
Project Summary

Summary

Agentshire transforms AI agents from abstract chat interfaces into living 3D NPCs within a customizable game town. This OpenClaw/QClaw plugin targets developers and power users seeking a persistent, visual environment for AI interaction, offering social simulation and user-generated content tools for a richer agent experience.

How It Works

The system integrates AI agents as 3D citizens in a simulated town, featuring dynamic day/night cycles, weather, and social interactions. It employs a dual-mode interface: a 3D "Town Mode" for observation and an "IM-style Chat Mode" for direct communication. Core components include a simulation layer, UGC tools like a map editor and character workshop, and an optional AI-driven "Soul Mode" for deeper agent personalities and conversations, all orchestrated via the OpenClaw runtime.

Quick Start & Requirements

  • Installation: OpenClaw CLI: openclaw plugins install agentshire. QClaw Desktop: Clone to extensions directory, npm install, restart QClaw.
  • Prerequisites: OpenClaw CLI 2026.3.13 (recommended) or QClaw Desktop 0.2.x. Node.js >= 18.
  • Resources: Optional asset pack is ~4.4GB.
  • Links: Demo video (asset link provided), VISION.md, ROADMAP.md.

Highlighted Details

  • Dual-Mode UI: Seamlessly switch between immersive 3D town visualization and a functional chat interface.
  • UGC Tools: Citizen Workshop for AI-assisted character creation and Town Editor for drag-and-drop map building.
  • Procedural Audio: Real-time synthesized ambient sounds and dynamic BGM without requiring audio files.
  • Agent Simulation: Algorithm-driven NPC behavior with an optional "Soul Mode" for LLM-powered multi-turn conversations and relationship tracking.

Maintenance & Community

No specific details on contributors, sponsorships, or community channels (Discord/Slack) are provided in the README. A roadmap is referenced via ROADMAP.md.

Licensing & Compatibility

  • License: MIT.
  • Compatibility: Recommended for OpenClaw CLI 2026.3.13 and QClaw Desktop 0.2.x. Compatibility with OpenClaw CLI 2026.4.x+ is explicitly noted as not yet supported due to an upstream regression.

Limitations & Caveats

OpenClaw CLI 2026.4.x+ is not supported due to an upstream channel initialization regression. Advanced visual assets require a substantial ~4.4GB download. The "AI Generate" feature for character souls necessitates a correctly configured LLM provider in openclaw.json.

Health Check
Last Commit

1 week ago

Responsiveness

Inactive

Pull Requests (30d)
1
Issues (30d)
1
Star History
337 stars in the last 30 days

Explore Similar Projects

Starred by Andrew Ng Andrew Ng(Founder of DeepLearning.AI; Cofounder of Coursera; Professor at Stanford), Jack Lukic Jack Lukic(Author of Semantic UI), and
5 more.

ag2 by ag2ai

0.4%
4k
AgentOS for building AI agents and facilitating multi-agent cooperation
Created 1 year ago
Updated 7 hours ago
Feedback? Help us improve.