A multi-agent AI system using Google AI Agent SDK and Gemini to analyze real-time route safety (crime, weather, lighting) and optimize navigation.

SafeRouteAI – Multi-Agent Real-Time Route Safety Advisor

A multi-agent AI system that analyzes route safety in real-time by combining crime data, weather conditions, traffic, time of day, and lighting conditions to provide safety recommendations.

🌟 Project Overview

SafeRouteAI is an autonomous multi-agent system built with Google AI Agent SDK that evaluates the safety of walking/driving routes by analyzing multiple risk factors and recommending safer alternatives when needed.

Key Features

• Multi-Agent System: 5 specialized agents working together
• Real-Time Analysis: Weather, crime, lighting, and time-based risk assessment
• Route Optimization: Suggests safer alternative routes
• Memory & Sessions: Remembers user preferences and route history
• Observability: Comprehensive logging and tracing
• Parallel Processing: Batch route analysis support

🏗️ System Architecture

Agent Architecture

1. Route Analysis Agent: Extracts route coordinates, distance, and waypoints
2. Safety Data Agent: Gathers weather, crime, lighting, and time-of-day data
3. Risk Scoring Agent: Computes combined risk scores (0-10 scale)
4. Route Optimization Agent: Suggests safer alternatives for high-risk routes
5. Alert Agent: Generates human-readable safety guidance

Multi-Agent Patterns

• Sequential Agents: Route Analysis → Safety Data → Risk Scoring → Alert
• Parallel Agents: Batch route analysis with concurrent processing
• Conditional Agents: Route Optimization only triggers for moderate/high risk routes
• Loop Agents: Session memory main