Agent Town

Agent Town is an innovative pixel-art AI agent collaboration platform that transforms AI orchestration into an engaging interactive experience. This unique online office environment allows users to walk through a vibrant pixelated workspace where AI agents actively complete tasks instead of relying on traditional dashboards or task queues.

Core Features

• In-world Task Assignment: Interact directly with workers by walking up and assigning tasks through an RPG-style menu. Experience a more hands-on way of managing workloads without the hassle of forms or dropdowns.
• Visible Task Execution: Keep track of tasks as they move through various stages: queued, returning, sending, running, and completed. Worker status is displayed in real-time, providing clear visibility and transparency into their workflow.
• Worker Autonomy: Witness idle workers as they roam the office, engaging with different elements like whiteboards and coffee stations. When ready, they return to their desks to focus on the assigned tasks. Busy workers will queue additional tasks instead of being overwhelmed.
• Session Management: Enjoy the flexibility of managing multiple sessions seamlessly with efficient switching, context metering, and configurations for worker names, roles, and sprites.

How It Functions

Interact with AI workers by following a simple sequence:

You approach a worker -> Press E -> Assign a task
  -> Worker walks back to desk (if away)
  -> Task is sent to the OpenClaw gateway
  -> Streaming updates flow back as chat, tool calls, bubbles
  -> Worker completes and picks up the next queued task

Technology Stack

• Application: Built with Next.js 16, React 19, and TypeScript for a robust and interactive user experience.
• Game Framework: Utilizes Phaser 3 alongside Tiled maps and pixel sprite sheets for the engaging visual layout of the office environment.
• Gateway Integration: Communicates with the OpenClaw gateway using a WebSocket proxy for real-time updates.
• State Management: Handles state through React context combined with a reducer and a typed event bus for organized communication.

Future Enhancements

Planned features include:

• A library scene that embodies long-term memory storage with a walkable layout for exploration.
• A workshop scene that facilitates skill and tool management through interactive physical stations.
• An expanded town map including a marketplace to integrate third-party skills and resources.
• More developed worker personalities and schedules for enhanced interaction.
• Improved onboarding processes for new users to ease their introduction to the platform.

Agent Town encourages contributions, particularly from individuals interested in gameplay design, level design, and creating user-friendly experiences tailored for AI workflows. Discover the future of AI team collaboration with Agent Town.