An AI-powered real-time disaster rescue coordination platform built with Solace Agent Mesh (SAM) — an open-source framework for building event-driven multi-agent AI systems.

The Disaster Rescue Coordination System is an event-driven multi-agent platform designed to optimize disaster response operations. By leveraging Solace Agent Mesh's asynchronous architecture, the system coordinates AI agents to handle victim communication, rescue team deployment, and resource allocation in real-time.

• 🗣️ Natural Language Victim Communication - AI-powered chat interface for gathering critical incident details
• 📊 Intelligent Triage & Prioritization - Automatic severity assessment and priority scoring
• 📦 Resource Management - Automated inventory tracking and equipment allocation
• 🗺️ Live Situational Awareness - Interactive maps with real-time victim and team locations
• 🔄 Event-Driven Architecture - Guaranteed message delivery and loose coupling via Solace PubSub+

• Communicates with victims via natural language
• Extracts location, severity, headcount, and hazard information
• Assigns priority scores (CRITICAL/MODERATE/STABLE)
• Delegates rescue tasks to Deployment Agent and Severity Agent

• Receives victim assignments from Orchestrator
• Calculates required equipment based on scenario type
• Determines personnel requirements
• Provides GPS navigation and route optimization

cd sam-project
cp .env.example .env

Edit .env with your credentials:

# Solace Broker (local development)
SOLACE_BROKER_URL=tcp://localhost:55555
SOLACE_BROKER_VPN=default
SOLACE_BROKER_USERNAME=admin
SOLACE_BROKER_PASSWORD=admin

# LLM Configuration
SAM_LLM_PROVIDER=openai
SAM_LLM_MODEL=gpt-4
SAM_LLM_API_KEY=sk-your-openai-key

# Or use Anthropic Claude
# SAM_LLM_PROVIDER=anthropic
# SAM_LLM_MODEL=claude-sonnet-4-20250514
# SAM_LLM_API_KEY=sk-ant-your-key

# Create virtual environment
python3 -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
pip install solace-agent-mesh

Deploy to AWS ECS Fargate using Terraform:

cd terraform/environments/prod

# Initialize Terraform
terraform init

# Review changes
terraform plan

# Deploy infrastructure
terraform apply

GitHub Actions automatically:

1. Tests - Runs unit tests on every PR
2. Builds - Creates Docker images for SAM and frontend
3. Pushes - Uploads images to AWS ECR
4. Deploys - Updates ECS services with new images

Trigger deployment:

git push origin main

• Solace for the Agent Mesh framework
• OpenAI for LLM capabilities
• Google ADK for agent runtime