GaiaAI – Real-Time Climate & Disaster Resilience Platform

GaiaAI: Real-time AI for disaster prediction, risk mitigation, and climate resilience, empowering communities to act before it’s too late

Comment

Splash screen
Assets deployment screen
Interactive Globe screen
GAIA communications screen
Grid screen
System settings screen
Grid detailed screen

Driven by the urgency of climate resilience and a love for sci-fi interfaces, I built GaiaAI. I wanted to harness Gemini’s multimodal power to create a tool where AI logic amplifies human intuition, turning global chaos into calculated, life-saving strategy.

I engineered GaiaAI to bridge human intuition and AI logic. Using Gemini’s Live API, I oversee global disasters on a real-time 3D globe, generate tactical video simulations, and deploy assets via voice. It allows me to predict, visualize, and neutralize threats instantly through a resilient, offline-capable command interface.

I engineered GaiaAI using React and TypeScript, integrating the full Google GenAI ecosystem. I utilized Gemini 3 for strategic reasoning, Veo for video simulation, and the Live API for real-time voice control. I built a custom physics engine and procedural math algorithms to render the 3D globe via native Browser APIs, creating a resilient, offline-capable command center.

I engineered GaiaAI using React and the full Gemini suite. The core is a custom TypeScript physics engine rendering 45,000 particles on a 2D canvas to emulate 3D globes. Integrating Gemini 3 and Veo was powerful, but my biggest challenge was optimizing the Live API’s audio synchronization and implementing a robust "offline simulation" layer. This ensures the command center functions autonomously even when the satellite uplink fails.

I am proudest of engineering a custom 3D planetary engine using raw TypeScript math and Canvas API, achieving 60fps performance without heavy libraries. Integrating the Gemini Live API was a technical triumph, enabling real-time, bidirectional voice command with complex tool execution. Additionally, I successfully built a robust "offline-first" architecture. By combining local procedural generation with sophisticated fallback heuristics, I ensured the command center remains fully operational and strategic even when the satellite uplink is severed.

I learned that balancing raw AI power with human agency is critical. Integrating Gemini 3 taught me that 'thinking' latency isn't a blocker but a UX opportunity—visualizing the AI's reasoning builds trust. Physically modeling 45,000 particles in the browser pushed my understanding of optimization, proving web technologies can handle military-grade visualization. Most importantly, I realized that true resilience requires an "offline-first" architecture; by simulating the AI's logic locally when disconnected, I ensured the system empowers the user even in the absolute worst-case scenarios.

I plan to evolve GaiaAI from a standalone command console into a global, decentralized resilience mesh. The next phase involves integrating real-time IoT sensor arrays directly into the physics engine to provide ground-truth data for Gemini's strategic reasoning. I will also develop an AR field interface, allowing responders to see Gemini's tactical overlays superimposed on their physical environment. Finally, I aim to implement multi-agent orchestration, where specialized Gemini instances coordinate logistics, medical, and engineering assets simultaneously across continental scales, creating a truly self-healing planetary infrastructure.

Built With

Updates

Ian Omwenga started this project — Jan 09, 2026 06:25 AM EST

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.