🔮 Bayesian AI Scenario Simulator

An interactive, zero-backend Cognitive Wargaming tool designed for analysts, strategists, and decision-makers. The engine simulates non-linear future scenarios (2nd and 3rd-order consequences) using Large Language Models (Gemini 3.5 Flash Lite) anchored in a Directed Acyclic Graph (DAG) architecture.

---

🎯 Vision & Purpose

Traditional forecasting often suffers from linear bias. The Bayesian AI Simulator forces users to explore probabilistic branches, unexpected trade-offs, and systemic sectorial impacts across customizable timelines.

It does not serve as a financial oracle, but as a Strategic Stress-Testing Tool to evaluate "What If" scenarios under strict probabilistic boundaries.

---

✨ Key Features

• BYOK Architecture (Bring Your Own Key): 100% Client-Side execution. Your Google Gemini API key is stored exclusively in your browser's Local Storage. No external database or intermediate backend.
• Timeline Alignment: Dynamic spatial calculation anchoring scenarios to an absolute year coordinate, allowing precise parallel branch inspection.
• Probabilistic Confidence Metrics:
  • Relative Confidence: Independent probability of occurrence based purely on the immediate parent scenario.
  • Absolute Confidence: Cumulative path-dependent probability calculated from Node 0 (Genesis).
• Multi-Branching Parallel Realities: Generate up to 3 mutually exclusive scenarios for the same target year in a single inference call.
• Target Year Override: Force projections to a specific future year while strictly preventing temporal paradoxes (past-year traversal).
• Incremental Deep Dive: Expand specific node insights without polluting the DAG layout.
• Data Lifecycle Management: Export and Import your entire simulation tree via standard .json format.
• Multi-Language Support: Fully operational in Italian, English, Spanish, and French.

---

🛠️ Tech Stack & Architecture

• Frontend: Vanilla HTML5, CSS3 (Glassmorphism UI), JavaScript (ES6+).
• Graph Engine: Vis-Network (Hierarchical Directed Layout).
• AI Model: Google Gemini 3.5 Flash Lite via REST API.
• Persistence: HTML5 Web Storage API (localStorage).

---

🚀 Getting Started

Local Execution

1. Clone this repository or download the source files.
2. Open index.html in any modern web browser (Chrome, Safari, Firefox).
3. Click on the 📖 Guida / Guide or Inizializza Motore button.
4. Provide your Google Gemini API Key (Get one at Google AI Studio).

Cloud Deployment (Vercel / Netlify)

Since the application is purely static and backend-less:

1. Import this repository into Vercel or Netlify.
2. Click Deploy (No environment variables or build settings required).

---

🔒 Security & Privacy

This application does not collect, track, or proxy any user data. All API calls are routed directly from your browser to Google's official endpoints (generativelanguage.googleapis.com).

---

📄 License

This project is open-source and available under the MIT License.