Skip to content

nabil24024004/Navigator-AI

Repository files navigation

Navigator AI 🧭

AI-powered emergency crisis navigator that reasons over photos/videos to provide immediate, actionable intervention plans.

Navigator AI bridges the critical gap between an emergency event and professional response. By leveraging Gemini 3 Pro, it analyzes visual scenes in seconds to detect hazards, classify urgency, and generate tactical step-by-step checklists.

Status AI Stack

✨ Key Features

  • 📸 Multimodal Analysis: Upload images or take photos directly. The system identifies objects, hazards, and environmental context.
  • 📝 Tactical Intervention Plans: Converts chaotic situations into clear, prioritized checklists with estimated times for every step.
  • 🧠 Reasoning Trace: A "Chain of Thought" UI that explains why the AI recommended specific actions, building user trust.
  • 🗣️ Live Voice Mode: Real-time, low-latency voice conversation using gemini-2.5-flash-native-audio for hands-free assistance.
  • 🚨 Urgency Classification: Automatic detection of Critical, Medium, or Low urgency states with appropriate UI alerts (visual/haptic/audio).
  • 💾 Offline-First: Saves incident history locally and caches the active plan for reliable access in spotty network conditions.

🚀 Getting Started

Prerequisites

  • Node.js 18+
  • A Google AI Studio API Key with access to gemini-3-pro-preview and gemini-2.5-flash-native-audio-preview-09-2025.

Installation

  1. Clone the repository

    git clone https://github.com/your-org/navigator-ai.git
    cd navigator-ai
  2. Install dependencies

    npm install
  3. Configure API Key Set your API key in your environment variables.

    export API_KEY="your_google_ai_studio_key"

    Note: In a local development environment using Vite/Webpack, you might need to use a .env file. Ensure process.env.API_KEY is accessible.

  4. Run the development server

    npm start
    # or
    npm run dev
  5. Open the app Navigate to http://localhost:3000 (or the port shown in your terminal).

📚 Documentation

Detailed documentation regarding the project roadmap, architecture, and AI implementation can be found in the docs/ folder:

🛠️ Tech Stack

  • Frontend: React 19, TypeScript, Tailwind CSS
  • Icons: Lucide React
  • AI SDK: @google/genai (Unified SDK)
  • Models:
    • Analysis: gemini-3-pro-preview
    • Live Voice: gemini-2.5-flash-native-audio-preview-09-2025

⚠️ Safety Disclaimer

Navigator AI is a guidance support tool. It is not a replacement for professional emergency services (911/999/112). In life-threatening situations, always contact local authorities immediately.

📄 License

MIT License. See LICENSE for details.

About

Bridging the "Golden Hour" gap in crisis response using Gemini 3 Pro’s multimodal reasoning.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors