Inspiration

Cities today face increasingly complex challenges, including traffic congestion, extreme weather events, energy inefficiencies, environmental concerns, and delayed emergency responses. While technology has enabled cities to collect vast amounts of data, most systems still operate in isolation, limiting their ability to make coordinated decisions.

We were inspired by the idea of creating a city that could think, predict, and respond proactively rather than simply reacting to incidents after they occur. The concept of Agentic AI and Digital Twins offered a powerful opportunity to reimagine how urban environments can become more intelligent, resilient, and citizen-centric. This led us to build MetaEarth AI—an autonomous platform where multiple AI agents collaborate to manage city operations in real time.

What it does

MetaEarth AI is an autonomous smart-city intelligence platform powered by multiple AI agents operating within a real-time digital twin environment.

The platform continuously monitors and analyzes urban data from different domains, including traffic management, weather conditions, emergency services, energy consumption, and environmental monitoring.

Specialized AI agents work together to:

• Predict potential disasters and high-risk situations

• Optimize traffic flow and reduce congestion

• Coordinate emergency response strategies

• Improve energy distribution and resource utilization

• Monitor environmental conditions and identify risk zones

• Provide city administrators with real-time recommendations and insights

A central City Governor Agent coordinates all decisions, ensuring that individual agents work together to achieve the best outcomes for the city.

How we built it

We designed MetaEarth AI using a multi-agent architecture inspired by real-world city management systems.

The platform consists of:

• Traffic Agent

• Disaster Prediction Agent

• Emergency Response Agent

• Energy Optimization Agent

• Environmental Monitoring Agent

• City Governor Agent

Each agent is responsible for a specific operational domain while continuously communicating with other agents to share information and coordinate decisions.

For the technology stack, we leveraged modern AI and cloud technologies, including:

• Large Language Models for reasoning and decision-making

• Agent orchestration frameworks for multi-agent collaboration

• Real-time event processing and simulation

• Cloud infrastructure powered by AWS

• Interactive dashboard with digital twin visualization

The result is a unified system capable of simulating and managing complex urban scenarios in real time.

Challenges we ran into

Building a collaborative multi-agent ecosystem presented several technical and design challenges.

One major challenge was coordinating communication between multiple agents while avoiding conflicting decisions. Since each agent focuses on a specific domain, ensuring that all agents aligned toward a common city-wide objective required careful orchestration.

Another challenge was designing realistic urban simulations that could effectively demonstrate the impact of autonomous decision-making within the limited timeframe of a hackathon.

We also faced challenges in balancing system complexity with usability, ensuring that the platform remained intuitive and understandable for users and judges.

Accomplishments that we're proud of

We successfully developed a working proof-of-concept for an autonomous city management platform powered by Agentic AI.

Key achievements include:

• Creating a collaborative multi-agent architecture

• Simulating real-world urban scenarios

• Building a real-time digital twin dashboard

• Demonstrating proactive decision-making instead of reactive responses

• Integrating multiple operational domains into a single intelligent ecosystem

Most importantly, we transformed a highly ambitious concept into a functional platform capable of showcasing the future of smart-city management.

What we learned

Throughout the development process, we gained valuable insights into the potential of Agentic AI beyond traditional chatbot applications.

We learned that:

• Multi-agent collaboration can solve complex problems more effectively than isolated AI systems

• Digital twins provide a powerful environment for simulation and decision support

• Real-time coordination is critical for large-scale systems such as cities

• Human-centered design is essential when building AI systems that influence real-world operations

This project expanded our understanding of how autonomous AI can be applied to infrastructure, sustainability, and public services.

What's next for MetaEarth

MetaEarth AI is designed as a foundation for future intelligent cities.

Our next steps include:

• Integrating live IoT and sensor data

• Expanding predictive analytics capabilities

• Incorporating satellite and weather intelligence

• Developing advanced emergency-response automation

• Enabling citizen-facing AI assistants

• Scaling the platform to support entire metropolitan regions

Our long-term vision is to create a global smart-city operating system where autonomous AI agents continuously collaborate to improve safety, sustainability, and quality of life for millions of people.

MetaEarth AI is not just a platform for managing cities—it is a step toward building cities that can understand, adapt, and evolve alongside the communities they serve.